Stock Assessment Report No. 03-02of the

Atlantic States Marine Fisheries Commission

Terms of Reference & Advisory Reportfor the Atlantic Croaker Stock Assessment Peer Review

October 2003

Working towards healthy, self-sustaining populations for all Atlantic coast fish speciesor successful restoration well in progress by the year 2015

Stock Assessment Report No. 03-02of the

Atlantic States Marine Fisheries Commission

Terms of Reference & Advisory Reportfor the Atlantic Croaker Stock Assessment Peer Review

Conducted onOctober 8 & 9, 2003

Raleigh, North Carolina

A publication of the Atlantic States Marine Fisheries Commission pursuant toNational Oceanic and Atmospheric Administration Award No.NA03NMF4740078.

i

Preface Summary of the Commission Peer Review Process The Stock Assessment Peer Review Process, adopted in October 1998 by the Atlantic States Marine Fisheries Commission, was developed to standardize the process of stock assessment reviews and validate the Commission’s stock assessments. The purpose of the peer review process is to: (1) ensure that stock assessments for all species managed by the Commission periodically undergo a formal peer review; (2) improve the quality of Commission stock assessments; (3) improve the credibility of the scientific basis for management; and (4) improve public understanding of fisheries stock assessments. The Commission stock assessment review process includes evaluation of input data, model development, model assumptions, scientific advice, and review of broad scientific issues, where appropriate. The Stock Assessment Peer Review Process report outlines four options for conducting a peer review of Commission managed species. These options are, in order of priority: 1. The Stock Assessment Workshop/Stock Assessment Review Committee

(SAW/SARC) conducted by the National Marine Fisheries Service (NMFS), Northeast Fisheries Science Center (NEFSC) or the Southeast Data and Assessment Review (SEDAR) conducted by the National Marine Fisheries Service (NMFS), Southeast Fisheries Science Center (SEFSC).

2. A Commission stock assessment review panel composed of 3-4 stock assessment

biologists (state, federal, university) will be formed for each review. The Commission review panel will include scientists from outside the range of the species to improve objectivity.

3. A formal review using the structure of existing organizations (i.e. American

Fisheries Society, International Council for Exploration of the Sea, or the National Academy of Sciences).

4. An internal review of the stock assessment conducted through the Commission’s

existing structure (i.e. Technical Committee, Stock Assessment Committee). Twice annually, the Commission’s Interstate Fisheries Management Program (ISFMP) Policy Board prioritizes all Commission managed species based on species Management Board advice and other prioritization criteria. The species with highest priority are assigned to a review process to be conducted in a timely manner. In November 2002, the Atlantic croaker stock assessment was prioritized for a SEDAR peer review. A review panel was convened of stock assessment biologists and representatives from the fishing community and non-government organizations. Panel members had expertise in Atlantic croaker life history and stock assessment methods. The SEDAR review for the Atlantic croaker stock assessment was conducted October 8-9, 2003 in Raleigh, North Carolina.

ii

Purpose of the Terms of Reference and Advisory Report The Terms of Reference and Advisory Report provides summary information concerning the Atlantic croaker stock assessment and results of the SEDAR review to evaluate the accuracy of the data and assessment methods for this species. Specific details of the assessment are documented in a supplemental report entitled Atlantic Croaker Stock Assessment Report for Peer Review. To obtain a copy of the supplemental report please contact the Commission at (202) 289-6400.

iii

Acknowledgments Thanks are due to the many individuals who contributed to the Commission’s Atlantic croaker Stock Assessment Peer Review. Special thanks are extended to the Atlantic Croaker Peer Review Panel (Dr. Steve Bobco, Old Dominion University, William Goldsborough, Chesapeake Bay Foundation, Najih Lazar, Rhode Island Division of Environmental Management Marine Fisheries Section, Dr. Tom Miller, Chesapeake Biological Laboratory, Dr. Jim Nance, NOAA Fisheries NMFS SEFSC, Dr. Paul Nitschke, NOAA Fisheries, NMFS NEFSC, Lee Paramore, North Carolina Division of Marine Fisheries, Dr. Stephen Smith, Bedford Institute of Oceanography, Dr. Elizabeth Wenner, South Carolina Department of Natural Resources, Geoffrey White, Atlantic States Marine Fisheries Commission, William T. Windley, Jr., Maryland Saltwater Sportfish Association) for their hard work in reviewing the meeting materials and providing advice on improvements to the Commission’s Atlantic croaker stock assessment. The Commission would like to extend its appreciation to the members of the Atlantic Croaker Technical Committee and Stock Assessment Subcommittee for development of the Atlantic Croaker Stock Assessment Report for Peer Review (Stock Assessment Peer Review Report 03-002 Supplement) and specifically to the following members for presenting this report at the Peer Review meeting: Dr. Janaka DeSilva (Florida Fish and Wildlife Commission), and Dr. Eric Williams (National Marine Fisheries Service, Beaufort Laboratory). Special appreciation is given to the staff dedicated to the performance of the peer review and finalization of peer review reports, specifically – Dr. Lisa Kline, Dr. John Merriner, and Nancy Wallace.

iv

Table of Contents Preface.............................................................................................................................................. i Acknowledgments ........................................................................................................................ iii List of Figures................................................................................................................................ v Terms of Reference for the Atlantic croaker Peer Review ....................................................... 1

1. Evaluate the adequacy and appropriateness of fishery-dependent and independent data used in the assessments (i.e. was the best available data used in the assessment). ........ 1

2. Evaluate the adequacy, appropriateness and application of models used to assess these species and to estimate population benchmarks. ............................................................ 2

3. Evaluate the adequacy and appropriateness of the Technical Committee’s recommendations of current stock status based on biological reference points. ............ 4

4. Develop recommendations for future research for improving data collection and the assessment....................................................................................................................... 4

Atlantic Croaker Advisory Report.............................................................................................. 7

Status of Stocks........................................................................................................................... 7 Stock Identification and Distribution.......................................................................................... 7 Management Unit........................................................................................................................ 7 Landings...................................................................................................................................... 7 Data and Assessment .................................................................................................................. 8 Biological Reference Points........................................................................................................ 8 Fishing Mortality ........................................................................................................................ 8 Recruitment................................................................................................................................. 9 Spawning Stock Biomass............................................................................................................ 9 Bycatch ....................................................................................................................................... 9 Sources of Information ............................................................................................................... 9

v

List of Figures Figure 1. Atlantic coastal commercial landings of Atlantic croaker (metric tons), 1950-2001... 10 Figure 2. Recreational Landings (Type A+B1 in numbers) of Atlantic croaker ......................... 10

1

Terms of Reference for the Atlantic croaker Peer Review

1. Evaluate the adequacy and appropriateness of fishery-dependent and independent

data used in the assessments (i.e. was the best available data used in the assessment). The Atlantic croaker stock assessment used commercial and recreational landings data, the National Marine Fisheries Service (NMFS) Northeast Fisheries Science Center (NEFSC) bottom trawl indices, Marine Recreational Fisheries Statistics Survey (MRFSS) CPUE index, and Southeast Area Monitoring and Assessment Program (SEAMAP) nearshore trawl survey indices. The commercial landings data used in the assessment did not include landings from aggregate, unculled (“scrap”) bait fisheries nor were discard data estimated. Unculled bait landings data are only available from North Carolina and indicated that this fishery could account for a substantial amount (2-50%) of additional landings not accounted for in the directed fishery landings, particularly prior to 1996. The Panel expressed concern both over whether unculled bait landings data are available from other states and the magnitude of these landings for other states. The Panel recommends that the North Carolina unculled bait fishery data be evaluated and the landings updated to include these landings. The possibility of applying the North Carolina proportions to other states to estimate their unculled fish landings should also be explored. The unculled bait fishery consists of primarily small fish compared to other commercial landings and may require a revised or new selectivity curve in the model. The Panel also recommends that at-sea observer data be evaluated for inclusion of discard/bycatch data in the model. The Panel agreed that the MRFSS recreational landings for the period 1981 to present were the best available data. The Panel noted that as the ratio of commercial to recreational landings for the period 1981-present was used to hindcast earlier recreational landings, changes in the commercial removals (see above) will require re-estimation of recreational landings for the 1973 to 1980 time period. The Panel agreed with the validity of the recreational landings and the method of extending these data back to 1973. The model used NMFS NEFSC fall bottom trawl survey indices from 1982 to the present. The survey is a stratified random survey design extending back to 1963. The assessment used a survey index derived from the application of a delta lognormal model to the NEFSC bottom trawl data, as opposed to stratified mean estimates. A comparison of the delta lognormal estimates with stratified mean estimates on assessment results indicated substantial differences. The Panel noted that these differences were not addressed in the assessment report and was not confident in the use of the Delta lognormal model. The Panel recommends that the time series be extended back to 1973 and an evaluation be conducted to better understand the differences between the lognormal and stratified mean estimates. The Panel accepted the SEAMAP nearshore trawl data and the MRFSS CPUE index as the best data available. The Panel accepted the definition of croaker trips as consisting of a suite of species.

2

The final stock assessment model did not include trawl survey data from the Virginia Institute of Marine Science (VIMS). The VIMS trawl survey is believed to reflect dynamics of young croaker. The Panel noted that although the inclusion of the VIMS trawl survey might not be appropriate in an unstructured surplus production model, the VIMS time series may provide important information for the current assessment model. The Panel recommends further investigation of the inclusion of the VIMS trawl survey in the model since this survey covers the full time period and areas not covered by other survey indices included in the model. The assessment model uses a growth curve derived only from North Carolina data and applies this growth curve to all areas included in the model. The Panel agrees that the North Carolina growth curve is the best available. However, the Panel expressed concern that the North Carolina growth parameters were being applied across the entire latitudinal range of the stock and over the entire period of the assessment. Given the wide latitudinal range of this species, and the wide range of abundances observed in the stock, the Panel recommends the investigation of spatial and interannual variability in growth. Several different methods of calculating natural mortality (M) were evaluated in the stock assessment. The model used a constant M of 0.3 from the mid-point of the range of estimates. The Panel accepted the approach for calculating M as the best available, but recommends the development of age-specific mortality estimates. The Atlantic croaker stock assessment is not an age based method. Work is currently being conducted to standardize ageing methods for Atlantic croaker. The Panel recommends that the Commission conduct an ageing workshop to develop approved standard ageing protocols to improve coastwide consistency in ageing data. The Panel also support continued collection of age samples from fisheries-independent surveys and length samples from the MRFSS in order to improve future Atlantic croaker assessments.

2. Evaluate the adequacy, appropriateness and application of models used to assess these species and to estimate population benchmarks.

The model used was a forward projection age-structured production model with the age structure generated by the model and not included as input data to the model. The model was run separately for the mid-Atlantic region (North Carolina and north) and the south Atlantic region (South Carolina to Florida). The regions were separated due to a lack of observations of older fish (age 3+) in the southern region and differences in the temporal patterns in fishery-independent survey indices in the southern area, which indicated that dynamics may be different in the two regions. The Technical Committee indicated that performance of the mid-Atlantic model was acceptable, whereas that of the south Atlantic model was not wholly acceptable. There was extensive Panel discussion of the justification and implications of the separation of croaker into two management units. One view suggested the separation reflected a recognition of a lack of knowledge regarding the dynamics of croaker in the southern part of the species range. An extension of this view implies that the separation reflects a “culling” of the data so that the strength of the signal in the mid-Atlantic is not masked by differences in indices in the southern portion of the range. An alternative view is that there is indeed some functional stock structure that underlies the decision to develop separate models. An extension of this view implies the

3

potential for different reference points in the two components. Overall, the Panel did not believe that data were available to support either view. The Panel recommends investigation of the distribution and movement of croaker by age and season, and a comparison of life history parameters over the full distribution of croaker to address these uncertainties and provide full justification for a spatially explicit model. The Panel recommends tagging (artificial tags or natural tags such as otolith microchemistry/genetics) studies be conducted to address the justification for regional assessments. The model for the Mid-Atlantic region used commercial and recreational landings from 1973 to the present, while the survey indices used in the model only extended back to 1982. The Panel expressed concerns with starting the model in 1973 with landings data only and not taking advantage of the available survey tuning indices. During the review, the Panel requested a comparison model run using the NMFS NEFSC bottom trawl survey data from 1973 to the present. This analysis provided some indication of differences in scale between the full series and the partial series used in the assessment. The Panel recommends re-running the model using the full series of NMFS NEFSC fall bottom survey data. The Panel also recommends the evaluation and possible inclusion of the VIMS trawl survey data. The base model assumed that the SSB in 1973 was equal to 0.75 SSB (virgin biomass) from the Beverton-Holt analysis. The Panel was concerned about the validity of this assumption. The Panel recommends that the assessment readdress this assumption once the full time series of survey data is included in the model. The model assumes that the fisheries-independent survey indices are more precise than the fisheries-dependent data and recruitment deviation estimates and, therefore, provided higher weights to these surveys. The Panel did not find compelling evidence to support the weightings applied. The Panel noted that these weighting factors may not be optimum and could strongly impact model results. The Panel recommended an exploration of the consequences of different weighting factors. The assessment included an age structured production model only. This required development of an algorithm to generate an age structure for the population. The Panel recommends a comparison of non-age assessment models, such as the Collie-Sissenwine catch-survey and a delay difference model, to understand the implications of this age structure on derived reference points and stock advice. The Panel accepted selectivity curves used for both commercial and fisheries-independent indices as the best available. The Panel recommends the evaluation of culling the larger fish out of the survey indices to better match the assumed selectivity. The Panel noted that the assessment model relies on a single renewal function – specifically a Beverton and Holt stock- recruit function. The Panel noted that there has been dramatic variation in croaker abundance over the time period. In weakfish, a related sciaenid fish, similar variation in abundance has induced density-dependent changes in fecundity. If similar biological changes, or environmentally induced changes to potential stock productivity have occurred in croaker, the

4

assumption of a constant renewal function may be questionable. The Panel recommended an evaluation of changes in maturity and fecundity within the stock.

3. Evaluate the adequacy and appropriateness of the Technical Committee’s recommendations of current stock status based on biological reference points.

The Atlantic Croaker Technical Committee had concerns with recommending and evaluating reference points for the south Atlantic model at this time. Given the lack of data to estimate movement between the two regions, and the poor model fits, estimates of Fmsy and SSBmsy for the South Atlantic may be incorrect. The Panel accepted this conclusion regarding the southern region. The benchmarks for the mid-Atlantic region listed in the stock assessment report were corrected as follows:

F threshold - Fmsy Biomass threshold - 0.7 SSBmsy F target – 0.75 Fmsy Biomass target – SSBmsy

These benchmarks are based on Restrepo et al. (1998) and are standard for other managed species. The Panel noted that these benchmarks are appropriate given the model. Stock status determination was only provided for the Mid-Atlantic region, with F2001 = 0.98 Fmsy, and SSB2001 = 1.76 SSBmsy. Based upon the recent trends in survey indices, many members of the Panel accepted that the stock was not overfished; however, full consensus was not reached. However, given the lack of precision associated with the F estimates and the problems noted earlier with the model and landings data the Panel could not determine if overfishing is occurring. The Panel recommends that if the high degree of uncertainty in current F’s continues, a more conservative target be evaluated so that management action to meet the target F may not place the stock in danger of simultaneously exceeding the limit F. Given the major concerns with the landings data and abundance indices used in the model, the Panel expressed concern with use of the current Atlantic croaker stock assessment for management purposes. The Panel recommends that the Atlantic Croaker Technical Committee resolve the issues in research recommendations 1-7 and update the assessment.

4. Develop recommendations for future research for improving data collection and the assessment.

The Panel recommends that the Atlantic Croaker Technical Committee resolve the issues in research recommendations 1-7 during the development of an updated assessment.

1. Issue: Commercial landings did not include all removals from the population. • Evaluate North Carolina unculled bait (“scrap”) fishery data and include in the

commercial landings.

5

• Evaluate the potential of applying the North Carolina unculled bait fishery data to other states.

• Consider at-sea observer data for discards and bycatch.

2. Issue: The model used catch data from 1973 to the present but tuning indices were only used from 1981 to the present. • Extend the NMFS NEFSC bottom trawl survey data to 1973 for inclusion in the

model. • Evaluate the difference between the Delta lognormal and stratified mean estimates

from NMFS NEFSC bottom trawl survey. • Evaluate the VIMS survey data for possible inclusion in the model.

3. Issue: The base model assumed that the SSB in 1973 was equal to 0.75 SSB (virgin

biomass) from the Beverton-Holt analysis. • Re-evaluate after inclusion of the full time series of NMFS NEFSC and VIMS trawl

survey data.

4. Issue: The model assumes that the fisheries-independent survey indices are more precise than the fisheries-dependent data and model recruitment estimates and, therefore, provided higher weights to these surveys. • Evaluate the consequences of alternative weighting schemes. • Provide detailed justification for the final choice of weighting scheme.

5. Issue: Separate models were developed for the mid-Atlantic (North Carolina and north)

and South Atlantic (South Carolina to Florida). • Investigate the distribution and movement of croaker by age and season. • Compare life history parameters over the full distribution of croaker.

6. Issue: The assessment included an age structured production model only. This required

development of an algorithm to generate an age structure for the population. • Compare non-age assessment models, such as the Collie-Sissenwine catch-survey

and a delay difference model, to understand the implications of this age structure on derived reference points and stock advice.

7. Issue: Determination of overfishing/overfished were based on point estimates only.

• Estimate the error distribution for current estimates of F, and reference points. • Determine whether, given error distributions determined above, target F and threshold

F could be distinguished from estimates derived from the assessment model. • Consider revising F target reference point relative to the previous bullet.

The following research recommendations are lower priority, long-term research issues. These recommendations will provide improvements to future assessments. 8. Issue: Separate models were developed for the mid-Atlantic (North Carolina and north)

and South Atlantic (South Carolina to Florida).

6

• Conduct tagging and otolith microchemistry studies to address the justification for regional assessments.

9. Issue: Difficult to understand what component of the population the surveys were

tracking. • Include maps of fishery and survey areas in future reports.

10. Issue: A single growth curve based on data from North Carolina was applied over all

years and for whole area. • Evaluate the applicability of the North Carolina growth curve to all areas (spatial

variability). • Investigate interannual variability in growth.

11. Issue: A single natural mortality estimate was used for all ages and years.

• Develop age-specific M for inclusion in the model.

12. Issue: Trends in the recruitment deviations may indicate temporal bias in the recruitment model. • Assess whether changes in potential population reproductive capacities have changed

by quantifying patterns in the maturity ogive and size- and age-dependent fecundity. • Assess whether density dependent shifts in age- or condition-dependent timing of age

at maturity have occurred as in other sciaenids. • Assess whether temporal patterns in recruitment slope or asymptote have occurred.

13. Issue: There are no standard protocols for ageing of Atlantic croaker. • Conduct a workshop to develop and approve ageing standards for Atlantic croaker. • Continue collection of coastwide age samples from fisheries-independent surveys and

length samples from the MRFSS.

14. Issue: Selectivity curves were used for both commercial and fisheries-independent indices. • Evaluate culling of the larger fish out of the survey indices to better match the

assumed selectivity.

7

Atlantic Croaker Advisory Report

Status of Stocks The Atlantic croaker stock status for the South Atlantic region is unknown at this time. The South Atlantic region makes up a relatively small component of the total stock biomass. Stock status determination in terms of overfishing is also unknown for the mid-Atlantic region. Given that the forward projection age-structured production model did not account for a likely significant source of removals by the scrap fishery along with questions on biomass indices noted in the full Peer Review Panel Terms of Reference Report, the Panel could not determine if overfishing is occurring. Based upon the recent trends in survey indices, many members of the Panel accepted that the stock was not overfished; however, full consensus was not reached. Stock Identification and Distribution Genetic studies indicate a single genetic stock of Atlantic croaker on the Atlantic coast and separate, weakly differentiated stocks in the Atlantic and Gulf of Mexico. Management Unit The management unit for Atlantic croaker is the entire Atlantic coast from Delaware to Florida. Landings Commercial landings for Atlantic croaker exhibited three periods of peak landings: 1955-1959, 1975-1980, and 1995 to the present (Figure 1). The highest landings were in 1977 at 13,532 mt. The current period of elevated landings is more than seven years. Low levels of harvest were evident during the 1960s and 1970s. The commercial harvest has been dominated by North Carolina and Virginia since 1950. The commercial landings data did not include landings from aggregate, unculled (“scrap”) bait fisheries or discard data. Unculled bait landings data are only available from North Carolina and indicated a substantial amount of additional landings not accounted for in the model (2-50%), particularly prior to 1996. There is uncertainty whether unculled bait landings data are available from other states and the magnitude of these landings. Recreational landings are from the National Marine Fisheries Service Marine Recreational Fishery Statistics Survey (MRFSS). From 1981-2002, recreational landings of Atlantic croaker (Type A+B1 in numbers) from Massachusetts through Florida have varied between 2.8 million fish (1981) and 13.2 million fish (2001), with landings showing a strong linear increase over this period (Figure 2). Average landings for the period 1981 – 1990 were 6.0 million fish, while more recent landings averaged 10.8 million fish. The increased landings in recent years have been at the northern range of the fishery (Massachusetts to North Carolina).

8

Data and Assessment The Atlantic croaker stock assessment used commercial landings from NOAA general canvas reports for all states, including the east coast of Florida. No data from the scrap fishery were included in the assessment model. No observer data were evaluated to quantify discards. Biological samples were from state surveys from North Carolina since 1982, Virginia since 1989, and limited age/weight data from Maryland since 1999. Recreational landings data from 1981 to the present were from the MRFSS. A fishery dependent survey index of the MRFSS CPUE index was also used in the assessment. Fishery independent surveys included the National Marine Fisheries Service (NMFS) Northeast Fisheries Science Center (NEFSC) fall bottom trawl indices from 1982 to the present, and Southeast Area Monitoring and Assessment Program (SEAMAP) nearshore trawl survey indices from 1989 to the present. The assessment model used a deterministic age-structured surplus production model to explain the population dynamics of Atlantic croaker, where the population in successive years was linked using a Beverton-Holt stock recruitment relationship. For modeling purposes, the Atlantic croaker population was divided into two geographic regions: mid-Atlantic (all states north of and including North Carolina) and south Atlantic (all states south of and including South Carolina). Biological Reference Points No biological reference points have been determined for the South Atlantic region. The benchmarks for the mid-Atlantic region listed in the stock assessment report were corrected as follows:

F threshold - Fmsy Biomass threshold - 0.7 SSBmsy F target – 0.75 Fmsy Biomass target – SSBmsy

These benchmarks are based on Restrepo et al. (1998) and are standard for other managed species. Fishing Mortality The lack of inclusion of the landings in the scrap fishery in the assessment implies that removals were not fully accounted for in the model. Consequently, this suggests that estimates of F produced in the model have unknown biases. Given the lack of inclusion of all removal and questions on biomass indices, the Panel did not accept the fishing mortality estimates provided in the Atlantic Croaker Stock Assessment Report for Peer Review (include publication number here).

9

Recruitment The lack of inclusion of the landings in the scrap fishery in the assessment implies that removals were not fully accounted for in the model. Consequently, this suggests that estimates of recruitment produced in the model have unknown biases. Given the lack of inclusion of all removal and questions on biomass indices, the Panel did not accept the recruitment estimates provided in the Atlantic Croaker Stock Assessment Report for Peer Review (include publication number here) and suggests that trends in recruitment estimated by the model should be interpreted in relative terms. Spawning Stock Biomass The lack of inclusion of the landings in the scrap fishery in the assessment implies that removals were not fully accounted for in the model. Consequently, this suggests that estimates of spawning stock biomass produced in the model have unknown biases. Given the lack of inclusion of all removals and questions on biomass indices, the Panel did not accept the spawning stock biomass estimates provided in the Atlantic Croaker Stock Assessment Report for Peer Review (include publication number here). Bycatch Bycatch and discard information was not included in this stock assessment for commercial fisheries. Recreational discards were accounted for in the assessment. Sources of Information Atlantic States Marine Fisheries Commission. 2003. Atlantic Menhaden Stock Assessment

Report for Peer Review. ASMFC Stock Assessment Peer Review Report No. 03-02 (Supplemental). Washington, DC.

Restrepo, V.R., G.G. Thompson, P.M. Mace, W.L. Gabriel, L.L. Low, A.D. MacCall, R.D.

Methot, J.E. Powers, B.L. Taylor, P.R. Wade, and J. F. Witzig. 1998. Technical guidance on the use of precautionary approaches to implementing National Standard 1 of the Magnuson-Stevens Fishery Conservation and Management Act. NOAA Tech. Memo. NMFS-F/SPO-31. 56 p.

10

Figure 1. Atlantic coastal commercial landings of Atlantic croaker (metric tons), 1950-2001.

Figure 2. Recreational Landings (Type A+B1 in numbers) of Atlantic croaker

0

2000

4000

6000

8000

10000

12000

14000

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

Year

Land

ings

(1,0

00's

of f

ish)

0

2000

4000

6000

8000

10000

12000

14000

16000

‘50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00

year

mt

Atlantic States Marine Fisheries Commission

Atlantic Croaker 2004 Stock Assessment Supplement

DRAFT

May 2004

Working towards healthy, self-sustaining populations for all Atlantic coast fish species or successful restoration well in progress by the year 2015

2

Table of Contents 1.0 Introduction ................................................................................................................. 4 2.0 Addressing Panel Points on Data Inputs. ................................................................. 4

2.1 Commercial landings did not include all removals from the population .................. 4 2.1.1. Evaluate North Carolina unculled bait (“scrap”) fishery data and include the commercial landings ............................................................................................. 4 2.1.2 Evaluate the potential of applying the North Carolina unculled bait fishery data to other states .............................................................................................................. 5 2.1.3 Consider at-sea observer data for discards in ocean gill nets and ocean trawl and investigate the potential for developing bycatch estimates in the North Carolina shrimp trawl fishery. ................................................................................................... 6

2.2 Tuning indices. .......................................................................................................... 7 2.2.1 Extend the NMFS NEFSC bottom trawl survey data to 1973 for inclusion in the model. .................................................................................................................... 7 2.2.2 Evaluate the difference between the delta lognormal and stratified mean estimates from NMFS NEFSC bottom trawl survey. ................................................. 7

2.3 Model Formulations .................................................................................................. 9 2.3.1 The base model assumed that the SSB in 1973 was equal to 0.75 SSB (virgin biomass) from the Beverton- Holt analysis. Re-evaluate after inclusion of the full time series of NMFS NEFSC and VIMS trawl survey ............................................... 9 2.3.2 Evaluate the consequences of alternative weighting schemes. The model assumes that the fisheries-independent survey indices are more precise than the fisheries-dependent data and model recruitment estimates and, therefore, provided higher weights to these surveys. ................................................................................. 9

3.0 Final Model Run ....................................................................................................... 10 4.0 Management .............................................................................................................. 11

4.1 Risk Analysis. ......................................................................................................... 11 4.1.1 Determination of overfishing/overfished were based on point estimates only.................................................................................................................................... 11 4.1.2 Estimate the error distribution for current estimates of F, and reference points.................................................................................................................................... 12

5.0 Conclusions ................................................................................................................ 13

5.1 Determine whether, given error distributions determined above, target F and threshold F could be distinguished from estimates derived from the assessment model........................................................................................................................................ 13 5.2 Consider revising F target reference point relative to the previous bullet .............. 13

APPENDIX A: Estimating the Scrap Component of Atlantic croaker on the Atlantic coast of the United States using data collected by the North Carolina Division of Marine Fisheries .......................................................................................... 14

3

APPENDIX B: Estimating Virginia’s Scrap Landings: Using Virginia field sampling data .................................................................................................................. 41 APPENDIX C: Estimates of annual Atlantic croaker bycatch in the North Carolina shrimp trawl fishery, 1973-2002, based on a simple fish: shrimp ratio approach .... 48 APPENDIX D: An Evaluation of the NEFSC observer data to estimate Atlantic croaker discards .............................................................................................................. 71 APPENDIX E: Using the NMFS-Woods Hole Fall Groundfish Survey off the East Coast of the United States: 1972-2002 to develop indices of abundance for Atlantic croaker ............................................................................................................................. 84 Appendix F: An Evaluation of Weighting the Likelihood terms in an Age Structured Production Model for Atlantic croaker ...................................................................... 116 Appendix G. A re-assessment of the status of the Atlantic croaker population in the Mid-Atlantic (New York to North Carolina). ............................................................. 138

Summary of Changes .................................................................................................. 138 Data changes ............................................................................................................... 139 Model changes ............................................................................................................ 140 Output/results .............................................................................................................. 142 Recommendations and findings .................................................................................. 148 Literature cited ............................................................................................................ 149

Appendix H: Status of stock identification of Atlantic croaker along the east coast on the U.S. ...................................................................................................................... 176

4

1.0 Introduction

In November 2002, the Atlantic croaker stock assessment was prioritized for a SEDAR peer review (ASMFC 2003; SEDAR report). A review panel was convened of stock assessment biologists and representatives from the fishing community and non-government organizations. Panel members had expertise in the Atlantic croaker life history and stock assessment methods. The SEDAR review for the Atlantic croaker stock assessment was conducted on October 8-9, 2003 in Raleigh, North Carolina. The Atlantic croaker stock status for the South Atlantic region is unknown at this time. The South Atlantic region makes up a relatively small component of the total stock biomass. Stock status determination in terms of overfishing was also unknown for the mid-Atlantic region at the time of the October peer review. Given that the forward projection age-structured model did not account for a likely significant source of removals by the scrap fishery and there were questions on biomass indices noted in the full Peer Review Panel Terms of Reference Report, the Panel could not determine if overfishing was occurring. Based on the recent trends in survey indices, many members of the Panel accepted that the stock was not overfished; however, full consensus was not reached (ASMFC 2003; SEDAR report). The Panel described in their report several major issues that required additional work by the Technical Committee (TC). There were seven short term issues the panel felt should be addressed to update the stock assessment. The South Atlantic State-Federal Fisheries Management Board directed the TC to address five of the short-term issues. These Five of these issues are presented in detail below. The other two issues; a coast wide versus regional stock assessment, and the exploration of additional models will be done at a later time. The TC has started looking into the issue of a coast wide versus a regional model through status of the stock identification (Appendix H). This issue will be addressed further at the time of the next benchmark assessment. The detailed descriptions below, and the updating of the assessment only refer to the mid-Atlantic model. The status of the South Atlantic stock remains unkown. 2.0 Addressing Panel Points on Data Inputs.

2.1 Commercial landings did not include all removals from the population (Section 5.1.3 of original stock assessment report, pg 21)

2.1.1. Evaluate North Carolina unculled bait (“scrap”) fishery data and include the commercial landings

For the revised assessment we have included the Atlantic croaker scrap estimates developed by the North Carolina Division of Marine Fisheries (NCDMF) from 1986-2002. For 1973-1985 we estimated North Carolina’s scrap landings using a number of

5

methods (see Appendix A). It was evident that no one method works for all gear types, owing to the quality of the available data. For example, developing estimates based on unclassified bait landings for ocean gill nets are low or do not exist, so the estimates are low or close to zero. The regression models described in Appendix A, can produce very high and variable estimates. There was little justification, with the exception of the pound net fishery to suggest that scrap has any correlation to landings. In general the ratio method based on bait or total unclassified landings may be the most suitable. The unclassified finfish landings are likely to incorporate Atlantic croaker scrap. However, the proportion of Atlantic croaker scrap in these landings is unknown. Ratios based on estimates of Atlantic croaker scrap (from NCDMF) to the unclassified finfish landings from 1986 through 1990 are likely the most representative of North Carolina scrap estimates for the 1973-1985 period. North Carolina scrap estimates for 1973-1985 were based on the average ratio of scrap to total unclassified finfish landings (1986-1990) and included in this assessment.

2.1.2 Evaluate the potential of applying the North Carolina unculled bait fishery data to other states

Using data from North Carolina, four approaches to estimating Virginia’s scrap landings were evaluated (see Appendix A for details). While estimates of Virginia’s scrap landings can be made using data from North Carolina. There is little ancillary data to evaluate the validity of those assumptions. Trip based methods may be the most appropriate; however, the time series of the number of trips by gear is limited. This method assumes that scrap estimates per trip for North Carolina and Virginia are similar. An alternate method to estimate Virginia’s scrap landings, using the bio-profile data collected by the Virginia Marine Resources Commission (VMRC) was also developed and evaluated by the technical committee (see Appendix B for details). The technical committee concluded that using the field sample of lengths from the Virginia harvest to estimate Virginia scrap was preferable to using data from North Carolina because there are distinct regional differences among the gear, area, and seasonal contributions to the Atlantic croaker landings and scrap. For example, the majority of the scrap in North Carolina stems from ocean trawl fisheries in coastal waters during late fall through winter, whereas the Virginia scrap primarily represents harvest from inside waters by pound net and haul seine fisheries during spring through late summer. The VMRC contacted long-time, high-volume seafood buyers (one on the western and one on the eastern shore) that wholesale Atlantic croaker from pound nets. The buyers indicated that Atlantic croaker less than 9 inches could generally be considered as scrap. However, both buyers and a middle peninsula buyer indicated that some small-size croaker (< 9 inches) was sold for food during years of low Atlantic croaker abundance. The buyers generally agreed that ½ of croaker within the 9-inch interval are sold as food fish, with a greater amount of this size category in the bait in recent years and less in earlier years. The technical committee endorsed using Atlantic croaker length data, collected by the VMRC, as the best method for estimating the scrap component of Atlantic croaker landings in Virginia.

6

2.1.3 Consider at-sea observer data for discards in ocean gill nets and ocean trawl and investigate the potential for developing bycatch estimates in the North Carolina shrimp trawl fishery. The technical committee evaluated the use of the Northeast Fisheries Science Center (NEFSC) observer database to estimate at-sea discards of Atlantic croaker in the gill net and trawl fisheries (see Appendix D for details) The group also investigated the potential for developing bycatch estimates for the North Carolina shrimp fishery (see Appendix C for detailed methods and results). For the at-sea discards, both ratio and trip based estimators were developed for the gill net and trawl fisheries. The regression approach, based on the log-log transformation, produced very low discard estimates. For the trip-based approach, effort information for the otter trawl fishery for Virginia was unavailable and it had to be assumed that the discard ratios observed in the coastal waters of the Atlantic ocean are applicable to the inshore gill net trips for Virginia. At best, trip based estimates can be estimated for the period 1993-2002, for all other periods an alternate approach would need to be used. Since the average number of trips sampled per year is low (< 25 trips), estimates based on yearly samples by gear are poor, A ratio-based method would use a consistent methodology to estimate the entire time series, but the correlation between landings and discards is weak. The technical committee endorsed using estimates based on the ratio of discards to landings in the final model. The technical committee also evaluated all available data on shrimp bycatch and made preliminary estimates of Atlantic croaker bycatch in the North Carolina shrimp fishery. Based on available size data, the majority of Atlantic croaker bycatch would be Age 0. Estimates of Atlantic croaker bycatch in the shrimp fishery are highly uncertain. The majority of data were collected in one year of the NMFS observer program (1994). Other available data sources were poor. The preliminary estimates of Atlantic croaker bycatch may not capture the inter-annual variability across the time series, as estimates for 1973-1991 are based on 39 tows, 1992-1998 on 685 tows and 1999-2002 on 56 tows (See Appendix C for details). By consensus, the technical committee had no confidence in the inter-annual magnitude of the shrimp bycatch, for an approximate estimate of 10 million pounds for the 1994-95 period,. A Monte Carlo simulation of using available data indicated a high variability in the estimates. While, the shrimp bycatch is likely to be an important source of mortality, there appears to be little data to evaluate its magnitude. The technical committee concluded further work needed to be carried out on estimating Atlantic croaker bycatch in the shrimp fishery and therefore did not include it in the assessment at this time. Evaluating the effectiveness of developing estimates of discards by combining available information on the effectiveness of bycatch reduction devices with estimates of the effective ‘swept area’ by the shrimp fishery and abundance estimates from the SEAMAP and NCDMF indices need to be explored in more detail. The shrimp fishery has undergone significant changes in efficiency with the introduction of bycatch reduction devices and turtle excluder devices. Given the potential magnitude of estimates known with reasonable confidence (1994), sensitivity of the biological

7

reference points to the inclusion/non-inclusion of estimates from shrimp bycatch was examined. 2.2 Tuning indices. (Section 6.2.1 of original stock assessment report, pg 38) 2.2.1 Extend the NMFS NEFSC bottom trawl survey data to 1973 for inclusion in the model.

The National Marine Fisheries Service (NMFS) NEFSC trawl survey was re-examined, and data from 1973 through 2002 was included in the revised model. In the re-analysis of the NEFSC trawl index, estimates were based on numbers, as the quality of weight data in the early part of the time series was poor (not always taken). A detailed analysis of the data set was carried out and annual estimates, based on the stratified means, were developed (CW-STRAT; see Appendix E for details). In addition, the data were used to develop estimates using the delta-lognormal distribution and compared to stratified mean estimates developed by NMFS (courtesy of P.Nitske, NEFSC). Correlation between the stratified mean estimates developed by NMFS and CW-STRAT were high (0.94) and both methods exhibited a similar trend (Table 2.2.2.1).

2.2.2 Evaluate the difference between the delta lognormal and stratified mean estimates from NMFS NEFSC bottom trawl survey.

Comparison of the delta-lognormal estimates to the NMFS and CW-STRAT indicated that estimates from the delta-lognormal method were not consistent with the estimates derived from the stratified means, with extremely high estimates associated with the delta-lognormal method, for the early part of the time series (Table 2.2.2.1). The delta-lognormal model, treated depth as a categorical variable consisting of five classes. Closer examination of the data revealed that poor sample sizes within the categories were the likely cause for differing results. Least square mean estimates using a General linear model, where the response variable was the log (number +1) and explanatory variables were year, stratum and water temperature, were also carried out. While the scale of the least square mean estimates differed from the stratified mean estimates, trends between the methods were similar, with the exception of estimates between 1991 and 1994 and those in 2002 where the least square mean estimates indicated lower than average estimates (Table 2.2.2.1). The technical committee evaluated the different methods and concluded that the stratified mean estimates (CW-Strat) were the most appropriate for use in the model. These estimates are only based on strata that were suitable Atlantic croaker habitat (see Appendix E for details).

8

Table 2.2.2.1. Estimates of mean number of Atlantic croaker/tow from the NEFSC-NMFS trawl survey based on stratified, least square and delta-lognormal means. CW = estimates based on using only strata considered Atlantic croaker habitat. NMFS = all strata suitable for comparison. STRAT= stratified mean estimates. LSM =Least square mean estimates from GLM. Delta = delta lognormal estimates. Year CW-Strat NMFS-Strat LSM-CW LSM-NMFS DEL-CW DEL-NMFS1973 38.07 40.98 4.52 7.63 1.38 1,158.681974 143.20 158.79 6.63 10.96 1.16 1,375.751975 638.21 792.47 18.95 36.12 12.60 6,767.311976 397.61 376.20 17.01 32.42 15.85 7,285.821977 119.35 116.99 7.28 12.50 2.48 1,237.611978 161.72 125.16 6.70 11.11 0.09 438.961979 15.64 15.64 2.73 4.17 0.06 157.431980 88.53 99.22 1.72 3.01 0.06 329.371981 31.77 42.82 1.67 2.72 0.02 104.191982 9.11 5.67 1.37 1.42 0.01 63.321983 231.94 337.63 3.70 6.13 0.10 621.321984 267.61 303.28 13.93 27.58 0.33 1,373.501985 213.97 237.34 8.06 16.67 5.59 2,574.771986 127.11 99.11 14.02 22.20 0.16 409.771987 111.96 156.77 1.54 2.63 0.03 173.951988 31.65 41.50 7.56 12.78 0.02 125.241989 99.64 142.55 7.39 13.26 0.06 329.241990 79.82 65.03 2.37 3.79 0.06 343.981991 260.53 315.41 3.35 5.18 0.06 453.061992 216.19 219.49 5.89 7.64 0.13 539.001993 140.88 90.27 4.10 6.08 0.04 334.141994 478.57 309.95 9.60 18.18 0.29 1,450.531995 189.36 212.36 8.96 19.22 0.34 1,015.761996 203.99 173.92 8.80 18.07 0.21 922.631997 159.14 134.91 4.29 6.75 0.09 464.581998 344.79 319.08 17.21 34.67 0.56 1,322.891999 734.45 685.19 34.54 84.92 2.41 3,238.182000 387.65 471.44 9.40 16.24 0.20 721.432001 177.64 162.09 8.61 10.30 0.12 448.112002 939.82 676.95 19.42 28.80 0.59 1,629.70

9

2.2.3. Evaluate the VIMS survey data for possible inclusion in the model.

The spring VIMS index was included in the revised model run. This index is a young of the year index and estimates are geometrical means in numbers. While this index is spatially limited to Chesapeake Bay, it extends across the time series (1973-2002). Preliminary analysis revealed that the pattern in recruitment deviations was closely associated with indices that had a strong age 0 component. The TC concluded that including the VIMS index into the revised model was beneficial, in that recruitment deviations would be more closely associated with the index and would improve the estimation of parameters in the Stock-Recruit relationship. Also, it improves data included earlier in the time series which reduces the overall variability.

2.3 Model Formulations 2.3.1 The base model assumed that the SSB in 1973 was equal to 0.75 SSB (virgin biomass) from the Beverton- Holt analysis. Re-evaluate after inclusion of the full time series of NMFS NEFSC and VIMS trawl survey

(Section 6.1.2 of the original stock assessment report, pg 33)

Preliminary analyses revealed that unless the model included abundance indices that covered the early part of the time series (~1973), the initial SSB: SSB virgin ratio was poorly estimated. This lead to deterministically fixing the ratio in the original version. In the revised model, two indices that cover the early part of the time series and enables the SSB 1973:SSB virgin ratio to be estimated by the model. 2.3.2 Evaluate the consequences of alternative weighting schemes. The model assumes that the fisheries-independent survey indices are more precise than the fisheries-dependent data and model recruitment estimates and, therefore, provided higher weights to these surveys. (Section 6.2.2 of the original stock assessment report, pg 41) In the original version of the age structured production model, we gave the fleets, recruitment deviations and MRFSS index a weight of λ =1 and all fishery independent indices a weight of λ=2. In this iteration of the model, we explore alternate weighting schemes in more detail. The likelihood components fall into three groups, the fleet, index and the recruitment deviation components. The fleet and index likelihood terms are based on the difference between the observed data and the predicted estimates. The recruitment deviation likelihood is based on differences from a mean of 0 (i.e. no deviation from the stock-recruit relationship). As such, weightings were treated in two groupings: 1.) weights for the fleets and indices and 2.) weights for the recruitment deviations. Weighting profiles for the fleets and indices were examined, while keeping the weight on the recruitment deviations constant at λ=1. All of the likelihood terms were estimated assuming a

10

lognormal distribution ( see Appendix F for detailed methods and results).Our examination of possible weighting options revealed a relatively flat response surface for the likelihood terms. It was evident that none of the weightings considered produced a fit better than the base model. Simulations indicated that increasing an individual weighting component (to > 5) produced relatively little reduction in the standard deviation of the residuals. There is no basis to objectively determine an appropriate weighting scheme. However, experience tells us that, the fishery independent indices should be given a higher weight than that of the fleets. The original weighting scheme appears to be a reasonable choice for the data (see Appendix F for details.) 3.0 Final Model Run In the revised model the following changes were made:

1) Estimates of North Carolina and Virginia’s scrap landings were included in the

model. A model where scrap estimates were treated as a separate component was chosen over one where scrap landings were included as part of the commercial landings.

2) Using data from the NEFSC observer database, estimates of at-sea discards for the gill net and otter trawl fishery have been included.

3) The NEFSC trawl survey index has been extended to the entire time series, and the stratified mean estimates in numbers were used.

4) The VIMS spring index has been included in the model. 5) The model now estimates initial SSB: SBB virgin ratio. 6) The selectivity patterns used for the fleets has been refined using selectivity

patterns estimated from an ‘un-tuned’ separable VPA by incorporating the length and age data for Virginia’s and North Carolina’s commercial fishery (1989-2002) and the recreational fishery’s size distribution (1981-2002).

7) Commercial landings for 2002 were updated.

Details of the major changes and results for the revised model are presented in Appendix G. For the base model, the steepness (0.76), natural mortality (0.3), growth and length weight relationships used were similar to those in the original version. In the revised model, fishing mortality rates (F) are based on the average population weighted F for ages 1-10+. Fishing mortality rates for Atlantic croaker exhibit a cyclical trend over the time series. From 1977 to 1979, F rose rapidly reaching a maximum of 0.5 in 1979. From 1980 onwards, F rapidly declined reaching its lowest levels in 1992 (Appendix G; Figure G3; Table G8). Since 1993, F has gradually increased and between 1997 and 2002 remained relatively stable at around 0.11. For the base mid-Atlantic run, the trend in population abundance indicates a step-wise increase reaching a peak of 974 million fish in 1999. Population estimates from 1999 to 2002 have ranged from 663 to 974 million fish. The number of Age-0 fish in the population exhibit a series of periodic recruitment spikes in 1975, 1983, 1991, 1998, and 2002. Between 1999 and 2002 the number of age-0 fish have ranged between 100-375 million fish. Spawning stock biomass estimates

11

(estimated as the proportion of mature females) exhibit a cyclical trend over the time series. From the early 1970’s to 1983 spawning stock biomass declined to its lowest level (11,746 MT). Between 1999 and 2002 spawning stock biomass estimates have ranged between 80-91,000 metric tons (See Appendix G for detailed report). Between 1973 and 2002 the relationship the different sources of removals has changed. In particular, estimates of scrap/discards reached their peak in 1979 (3,200 MT) and since then declined to their lowest levels in 2002 (425 MT). Between 1973 and 1995, scrap/discard removals averaged 1,687 MT per year, whereas between 1996-2002 scrap/discards averaged 595 MT per year. It appears that the significant reduction in removals of predominantly age-1 and younger fish may have contributed to relatively stable fishing mortality and spawning stock biomass estimates since the mid 1990’s. 4.0 Management

4.1 Risk Analysis.

4.1.1 Determination of overfishing/overfished were based on point estimates only.

Burnham and Anderson (1998) define precision as “ a property of an estimator related to the amount of variation among estimates from repeated samples”. The model developed in excel, does not provide any estimates of precision. For models run using AD model builder, estimates of standard deviation are based on the delta method, which approximate the variance estimates. Variance estimates using the delta method are biased to the lower range of the spectrum when additional constraints are imposed on the model. Confidence bounds on the parameters can be estimated using bootstrap procedures. However, the estimates derived are likely to be biased (Hilborn and Walters, 1992). Ideally, the relative levels of confidence of the parameter estimates should be evaluated using methodology such as the “operating model concept” described in Hilborn and Walters (1992) or Bayesian methods; These are part of the long-term objectives in the model’s development.

As an interim measure, uncertainty in the estimates of the status of the stock was examined at three levels through a series of simulations. These were: 1) the sensitivity of the base model to alternate weightings of the likelihood components; 2) sensitivity of the model to alternative steepness and natural mortality estimates based on a prior distribution and 3) the implications of not including shrimp bycatch estimates (See Appendix G for details).

12

4.1.2 Estimate the error distribution for current estimates of F, and reference points.

For both fishing mortality and spawning stock estimates, estimates determined from the base run appear to be more pessimistic (conservative) when compared to other potential weighting schemes. This assumes that 3,500 simulations capture a wide range of weightings. The inter quartile range (25-75th percentile) for F2002 from the simulations ranged from 0.015 to 0.11 (See Appendix G for details). For 2002, average fishing mortality rates from the base model was close to the 75th percentile of the simulation runs (average F=0.11). The inter quartile range for 2002 spawning stock biomass estimates from the simulation ranged between 71,000 and 120,000 MT. In comparison, estimates of spawning stock biomass in 2002 from the base model was 80,000 MT, close to the value of 25th percentile of the simulation runs. Trends in fishing mortality and spawning stock biomass under varying steepness and natural mortality rates indicate that for 2002, the inter quartile range of spawning stock biomass estimates was 80,000 and 110,000 metric tons and between 0.08 and 0.12 for F2002. Based on the sensitivity runs, it appears that ~25% of the runs had higher fishing mortality estimates than those for the base run and ~25% of the sensitivity runs had spawning stock biomass estimates lower than the base run. Sensitivity analysis examining the inclusion of shrimp bycatch, indicated average fishing mortality rates (ages 0-10+) in 2002 ranged from 0.06 to 0.176 with 50% of the simulations having values less than 0.105. Spawning stock biomass estimates in 2002 from the simulation runs ranged from 77,000 to 149,000 MT with 50% of the values being less than 111,388 MT . In comparison, the average fishing mortality rate from the base run in 2002 was 0.11 (ages 1-10+) and the spawning stock biomass estimate in 2002 was 80,328 MT. Differences in Spawning stock biomass estimates are most likely a result of the model accounting for the increased removals as part of the shrimp bycatch by increasing the population estimates. (See Appendix G for details). Estimates of Fmsy from the base mid-Atlantic model was 0.39 and SSBmsy was equal to 28,932 MT. Estimates of average fishing mortality rates from the base mid-Atlantic model of 0.11 indicate that 2002 estimates were below the target and threshold levels (Appendix G). Recent estimates of SSB (~80,000 MT) are above both the proposed target and threshold levels. For 2002, F:Fmsy ratio was 0.263 and SSB:SSBmsy ratio 2.78. Based on the base run’s sensitivity to weighting of the likelihood components, and the sensitivity of the model to alternate steepness and natural mortality estimates, estimates derived from the base run appear robust. From the sensitivity analysis on weighting of the likelihood terms, 90 % of the simulations had F2002:Fmsy ratios less than 0.44 ( Appendix G). Biomass reference points from the weighting analysis indicated that 10% of the runs had SSB2002: SSBmsy ratios less than 2.27 (Appendix G). Model sensitivity to steepness and natural mortality estimates also indicated the stock was most likely below the fishing mortality targets and thresholds and above the biomass targets and thresholds; 90 % of the simulations had F2002:Fmsy ratios less than 0.44 and 10% of the runs had SSB2002:

13

SSBmsy ratios less than 2.16. When including estimates of Atlantic croaker caught as shrimp bycatch, simulations revealed that the current status of the stock was similar to the base run where shrimp bycatch were not included; the stock is not overfished or undergoing overfishing. However, biomass reference points from the simulation runs indicated higher SSBmsy values and the lower estimates of SSB2002:SSBmsy than those obtained for the base model. The range of estimates for Fmsy (~0.4) was similar to the base model (~ 0.39). SSBmsy estimates from the simulation (ranged from 48,000-67,000 MT with a median of 56,467 MT) and were much higher than those for the base run (28,932 MT). The ratio of F2002:Fmsy ranged from 0.14-0.43 with 50% of the runs having estimates below 0.26. In comparison F2002:Fmsy from the base model was 0.263 (based on ages 1-10+). The ratio of SSB2002:SSBmsy for the simulations ranged from 1.55 to 2.27, with 50% of the runs having estimates less than 1.98.

5.0 Conclusions 5.1 Determine whether, given error distributions determined above, target F and threshold F could be distinguished from estimates derived from the assessment model. While this analysis does not capture all of the sources of uncertainty, examination of the effects of alternate weightings of the likelihood components and alternate steepness and natural mortality estimates indicate that reference points derived from the base run are robust, and suggest that there was less than a 10% chance that the population is overfished or undergoing overfishing. Sensitivity analysis evaluating the inclusion/non-inclusion of shrimp bycatch estimates, indicate that SSBmsy estimates are sensitive to the inclusion of Atlantic croaker caught as shrimp bycatch. However, increased SSBmsy estimates are also accompanied by higher SSB estimates. The ratio of SSB2002:SSBmsy when preliminary estimates of shrimp bycatch is included indicates that the stock is unlikely to be below the threshold estimates. 5.2 Consider revising F target reference point relative to the previous bullet Based on the simulation analysis, there appears little need to revise the F target reference points. Of concern, would be management goals that define biomass reference points in absolute terms. Differences in Spawning stock biomass estimates are most likely a result of the model accounting for the increased removals as part of the shrimp bycatch by increasing the population estimates. There appears to be some justification for revising the reference points for the biomass target and threshold to relative terms until a more comprehensive evaluation of Atlantic croaker from shrimp bycatch can be carried out. Alternatively, we could use the biomass reference points from the analysis, with the understanding that they are based on a model that does not include shrimp bycatch.

14

APPENDIX A: Estimating the Scrap Component of Atlantic croaker on the Atlantic coast of the United States using data collected by the North Carolina Division of Marine Fisheries Scrap landings of Atlantic croaker primarily occur in North Carolina and Virginia and are not accounted for in the NMFS commercial landings database. Scrap landings of Atlantic croaker are primarily small fish that are part of the landings and used for bait or animal food. The North Carolina Division of Marine Fisheries (NCDMF) has developed scrap estimates of Atlantic croaker by gear type for the period 1986-2002 (NCDMF 2003). Scrap estimates of Atlantic croaker from Virginia are non-existent. This report summarizes the available data on scrap landings and the results of using that information to:

1) Estimate scrap landings of Atlantic croaker for North Carolina from 1973 to 1985.

2) Estimate Atlantic croaker scrap landings for Virginia from 1973 to 2002 using data from North Carolina.

Data Sources from North Carolina Since 1986, the state of North Carolina has estimated scrap landings of Atlantic croaker for the major commercial fisheries (NCDMF 2003). The North Carolina Division of Marine Fisheries (NCDMF) determined estimates of scrapfish landings for Atlantic croaker by applying the tri-annual ratio of marketable fish to scrapfish in the fish house samples to the reported tri-annual marketable landings (NCDMF 2003). In addition to Atlantic croaker scrap estimates, the number of trips and landings of Atlantic croaker by gear were available from the North Carolina trip ticket database from 1994-2002. Data Sources from Virginia Scrap estimates of Atlantic croaker from Virginia are non-existent. However, from 1993 to 2002, the annual number of trips and landings of Atlantic croaker by gear were available from the Virginia trip ticket database. The size distribution of Atlantic croaker by gear, was available from 1989-2002 from the Virginia Marine Resources Commission sampling program.

15

Other Data Sources While there are no estimates of scrap landings of Atlantic croaker in the NMFS commercial landings database, it does contain records for unclassified finfish by state and category. For state, year and gear specific information, data were available for four categories:

1) Unclassified finfish used as food 2) Unclassified finfish classified under the “general” category 3) Unclassified finfish spawn (fish roe; limited data for VA only) according to

NMFS roe estimates . 4) Unclassified finfish used as bait or animal food.

Estimates of unclassified finfish most likely include a diverse range of species and the proportions of these categories that represent Atlantic croaker are unknown in recent times. Historic information on the composition of the North Carolina scrap landings from 1962-64 indicated that Atlantic croaker comprised between 30-42% by weight of scrap from the trawl fishery and less than 1% by weight from the pound and haul seine fisheries (Fahy, 1966). The unclassified finfish estimates from the NMFS database could be considered as the upper bound of Atlantic croaker accounted for by scrap landings in North Carolina and Virginia. Any scrap estimates of Atlantic croaker would be expected to be less than the annual estimate of unclassified finfish in the NMFS commercial landings database. Summary of North Carolina Scrap Landings 1986-2002 For North Carolina, estimates of the scrap component of Atlantic croaker landings were available from 1986 to 2002 (NCDMF 2003). Estimates of Atlantic croaker in the North Carolina scrap landings were highest between 1987-1990, ranging from 1,249 to 1,569 metric tons and equivalent to about 50% of annual Atlantic croaker landings. More recently, North Carolina’s scrap landings have decreased as a result of a suite of regulations that mandated bycatch reduction devices in shrimp trawls (1992), eliminated fly net fishing south of Cape Hatteras (1994) and introduced culling panels in long haul seines (1999). , Estimates of scrap landings of Atlantic croaker from the North Carolina averaged 266 metric tons annually between 1997 and 2002, and were equivalent to 3% of that state’s Atlantic croaker landings (Table A1). In North Carolina, the primary gears that produced scrap landings between 1986-1993 were haul seines, ocean trawls (flynets), and pound nets. More recently, greater than 90% of scrap landings were from the haul seine and ocean trawls fisheries (Table A2). Scrap estimates for ocean otter trawls are a composite of three trawl types; (1) Flounder trawls (Otter trawl-Bottom, Fish);(2) Shrimp trawls and; (3) Flynets (Otter trawl-Midwater). Flynets are responsible for the majority of the scrapfish component in ocean trawls brought to the dock and sold as scrapfish. Shrimp trawls mainly cull at sea finfish too small for usual market grades.

16

Estimating Atlantic croaker Scrap Landings in North Carolina and Virginia North Carolina For the period 1986-2002, Pearson correlation estimates between the scrap estimates and potential explanatory variables were developed by gear type (Table A3). Preliminary analyses revealed that for most of the gears, the unclassified finfish used for bait and animal feed were more closely correlated to scrap landings than total Atlantic croaker landings (Table A3). For the gill net fishery there was poor correlation between scrap landings and any of the explanatory variables examined. To estimate North Carolina’s Atlantic croaker scrap component from 1973 through 1985, estimates using three different methods were explored.

1) A stepwise regression approach was used. The response variable was

the estimated scrap landings in pounds and explanatory variables were number of estimated trips (market landings *1,000 /cpue), Atlantic croaker landings, landings of unclassified finfish in the bait and animal food, unclassified finfish in the “general” category and unclassified finfish in the “food” category.

2) A linear regression approach where the log (scrap landings) was modeled, using either Atlantic croaker landings or the unclassified finfish landings for bait and animal food as explanatory variables. Using the parameter estimates and their associated standard errors, a Monte-Carlo simulation with 1,000 replicates was carried out to evaluate the error surrounding scrap estimates.

3) Using the last five years of available data (1986-1990) ratios of scrap landed to total croaker landings and scrap landed to total unclassified finfish used for bait and animal food and total unclassified finfish landings were developed for each gear type. These estimates were then used to apportion historic landings. Using the standard deviation estimates associated with these ratios, a Monte Carlo approach with 1,000 replications was carried out to estimate the error surrounding those estimates. A five-year interval was chosen to represent historic conditions, as it was the longest period where conditions were most likely to have been stable.

Stepwise Regression model Prior to developing the appropriate model for each gear, a Box-Cox transformation was carried out to determine the appropriate transformation of the response variable. With the exception of ocean gill nets, the log transformation of the response variable was appropriate (λ= 0). For ocean gill nets an appropriate λ value was –1.0, equivalent to 1/response variable.

17

To determine the most appropriate suite of explanatory variables that best predicted the scrap estimates by gear type, a stepwise regression was employed using liberal p-values to determine inclusion into the final model (entry into model p=0.4 and to be kept p=0.1). For the ocean gill net fishery a suitable regression model that met the minimum criteria could not be developed. For the otter trawl, pound net and haul seine fisheries adequate regression models that had R2 values > 0.5 were developed (Tables A4 and A5). Examination of the variables included in the final model indicated that for most gears, the number of trips was an important factor. However, data on the number of trips, for years where scrap landings need to be estimated, were unavailable and suitable proxies for effort are unavailable (NCDMF, personal communication). As such, the model was of little utility in estimating those missing years of scrap landings. Furthermore, for some of the other explanatory variables in the regression models, missing data limited their utility as a tool for estimating scrap landings for the missing years.

Given the limitations in using the stepwise regression model as a predictor, a simpler approach to estimating scrap landings was evaluated (methods 2 and 3).

Other Potential methods A generalized additive model (GAM) using both LOESS smoothers and regression splines on the explanatory variable was briefly examined. For each gear type, log (scrap landed in pounds) was modeled using either Atlantic croaker landings or total unclassified finfish landings as explanatory variables. Preliminary evaluations revealed that the GAM approach added little to treating the Atlantic croaker landings as a linear function and also had a poor fit. As a possible technique to estimate historical scrap landings for North Carolina multiple imputation methods were also examined. Multiple imputation is a Monte-Carlo technique in which missing values are replaced by M >1 simulated versions that represent the uncertainty about the correct value to impute. In this exploration, only the first phase of the imputation process, generating 20 complete data sets using the MCMC method was carried out. Analyses were performed using Proc MI in SAS. While multiple imputation techniques hold promise, further work needs to be carried out before applying these methods to estimating scrap landings. In the preliminary analyses, the posterior covariance matrix was singular and further work needs to be carried out in providing appropriate priors. Estimates of North Carolina Scrap Landings 1973-1985

Table A6 summarizes the unclassified finfish landings by category and gear for North Carolina. Table A7 summarizes the scrap estimates of Atlantic croaker using the regression and ratio methods. It was evident that no one method works for all gear types. In part this has to do with the quality of the available data. For example, unclassified bait

18

landings for ocean gill nets are low or do not exist, so the estimates turn out to be low or close to zero.

In general the ratio method based on bait or total unclassified landings may be the most suitable. The unclassified finfish landings are likely to incorporate Atlantic croaker scrap. However, the proportion of Atlantic croaker scrap in these landings is unknown. Ratios based on estimates of Atlantic croaker scrap (from NCDMF) to the unclassified finfish landings between 1986-1990 are most likely to be the most representative of North Carolina scrap estimates between 1973-1985. The regression models can produce very high and variable estimates. There is little justification, with the exception of the pound net fishery, to suggest that scrap has any correlation to landings. This is most likely due to regulations imposed in 1991, which limits the scrapfish catch to 5,000 pound per vessel per day.

Virginia Given the lack of any data for scrap landings in Virginia, using the NC data to estimate scrap landings from Virginia was examined. To estimate Virginia’s scrap landings from 1973 to 2002, we have used four approaches:

1) Use the parameter estimates from the stepwise regression (Table A4 and A5) to estimate Virginia’s scrap landings. For some years and gear, data were not available, and estimates were not produced.

2) Apply the regression estimators developed using method 2 for North Carolina

3) Apply the ratio estimators developed using method 3 for North Carolina

4) Develop estimates, using the estimated Atlantic scrap per trip from the NCDMF study, and number of trips for the gear from the VA trip ticket program. Also, this method has no SE estimates and only included 1993-2002 data.

Estimates of Standard errors were produced using 1,000 Monte-Carlo trials as described earlier. Estimates of Virginia Scrap Landings 1973-2002 Table A8 summarizes the unclassified finfish landings by category and gear for Virginia. Also included in Table A8 are Atlantic croaker landings by gear, trips, and the NCDMF scrap catch rate by year (in kgs). Table A9 summarizes the scrap estimates and their standard errors for Virginia using the different methods. While estimates of Virginia’s scrap landings can be made using data from North Carolina, there is little ancillary data to evaluate the validity of those assumptions. Trip based methods may be the most appropriate; however, the extent of the time series where

19

the number of trips by gear exists is limited. This method also assumes that scrap estimates per trip for North Carolina and Virginia are similar. An alternate method to estimate Virginia’s scrap landings, using the bio-profile data collected by the Virginia Marine Resources Commission was also developed (see Appendix B for details of method and results). Literature Cited Fahy, W.E. 1966. Species composition of the North Carolina industrial fish fishery. Commercial Fisheries Review 28(7): 1-8. North Carolina Division of Marine Fisheries (NCDMF). 2004. Assessment of North Carolina Commercial Finfisheries, 2000-2002. Final Performance report for Award Number NA 76 FI 0286, 1-3. North Carolina Department of Environment and Natural Resources. North Carolina Division of Marine Fisheries.

20

Table A1. Estimates of bait (scrap) landings of Atlantic croaker from North Carolina for all fisheries combined. Source: NCDMF Total Market Bait (Scrap) Fishery

Year

Collections sampled

Landed (mt)

CPUE (kgs)

Landed (mt)

CPUE (kgs)

Landed (mt)

CPUE (kgs)

Fisheries 1986 333 4,598 1,601 4,033 1,318 565 283 combined 1987 338 4,280 1,208 2,995 886 1,286 321

1988 366 5,066 1,414 2,601 905 1,465 509 1989 327 4,469 1,382 2,900 843 1,569 538 1990 346 3,685 1,564 2,436 1,020 1,249 545 1991 381 2,456 611 1,462 372 992 239 1992 407 1,872 1,410 1,183 995 689 415 1993 273 1,996 4,451 1,428 3,428 527 1,022 1994 204 2,924 9,871 2,026 8,462 899 1,409 1995 193 3,800 7,615 2,643 6,659 1,157 956 1996 253 4,840 8,547 4,411 7,841 476 706 1997 229 5,146 11,623 4,802 11,207 344 416 1998 207 5,020 4,787 4,845 4,660 175 127 1999 256 4,953 13,118 4,559 12,793 394 325 2000 302 4,845 12,187 4,543 11,885 301 302 2001 299 5,593 8,951 5,375 8,620 218 332 2002 259 4,722 12,178 4,559 11,781 163 398

21

Table A2. Estimates of bait (scrap) landings of Atlantic croaker from North Carolina by gear. Source: NCDMF. Total Market Bait (Scrap)

Fishery Year

Collections sampled

Landed (mt)

CPUE (kgs)

Landed (mt)

CPUE (kgs)

Landed (mt)

CPUE (kgs)

Long haul 1986 176 1,690 2,653 1,392 2,168 298 485 1987 119 1,528 1,197 691 863 838 335 1988 169 1,777 1,496 1,177 1,118 600 377 1989 139 2,324 1,067 1,427 664 897 402 1990 147 2,649 1,990 1,769 1,327 880 664 1991 140 1,391 833 896 555 495 278 1992 155 490 301 423 267 66 34 1993 105 427 499 196 331 232 167 1994 65 680 649 47 58 633 591 1995 53 897 559 74 117 823 442 1996 85 244 626 163 443 81 183 1997 71 169 220 28 40 142 180 1998 70 82 190 11 45 70 145 1999 64 114 151 3 4 111 147 2000 61 231 270 31 55 200 215 2001 52 155 1,423 45 390 110 1,033 2002 62 67 382 14 81 53 301

Sciaenid 1986 57 368 681 233 431 135 250 pound net 1987 59 666 667 499 533 167 135

1988 54 733 358 466 235 267 124 1989 53 505 543 266 289 239 254 1990 61 420 598 220 306 200 291 1991 59 207 318 81 123 126 195 1992 43 52 80 14 20 38 60 1993 33 90 38 7 3 83 35 1994 22 9 100 3 43 7 57 1995 53 15 12 6 4 9 8 1996 33 100 18 5 1 95 17 1997 26 6 11 <1 1 6 11 1998 36 2 25 <1 <1 2 25 1999 41 202 132 6 1 196 132 2000 18 17 126 <1 <1 16 126 2001 16 13 173 11 139 2 35 2002 11 11 21 <1 <1 11 21

22

Table A2.. Continued. Total Market Bait (Scrap)

Fishery

Year

Collections sampled

Landed (mt)

CPUE (kgs)

Landed (mt)

CPUE (kgs)

Landed (mt)

CPUE (kgs)

Ocean 1986 75 1,415 22 1,415 22 <1 -sink net 1987 113 1,082 32 1,082 32 <1 -

1988 94 1,110 107 1,110 107 <1 - 1989 92 585 68 585 68 <1 <1 1990 90 305 10 292 9 13 <1 1991 136 356 21 349 21 6 <1 1992 155 428 56 422 55 7 10 1993 76 354 113 354 113 <1 <1 1994 79 622 314 622 314 <1 1 1995 68 872 242 872 242 <1 <1 1996 95 1,859 541 1,859 541 <1 <1 1997 71 1,274 370 1,274 370 <1 <1 1998 69 2,547 713 2,544 713 3 1 1999 122 1,770 402 1,770 402 <1 <1 2000 182 1,734 306 1,726 305 8 1 2001 190 2,372 525 2,372 525 <1 <1 2002 160 1,911 325 1,909 325 2 <1

Ocean 1986 25 1,125 4,612 993 3,794 132 818 trawl (fish) 1987 47 1,004 4,200 723 3,020 281 1,181

1988 49 1,446 3,606 848 2,244 598 1,363 1989 43 1,055 3,672 622 2,205 433 1,467 1990 48 311 942 155 399 156 542 1991 46 502 80 136 205 365 597 1992 54 902 4,638 324 3,184 578 1,455 1993 59 1,125 7,134 871 5,496 234 1,637 1994 38 1,613 14,599 1,354 12,512 259 2,087 1995 19 2,016 11,750 1,691 10,275 325 1,475 1996 40 2,639 15,485 2,337 14,180 302 1,305 1997 61 3,697 15,810 3,500 15,241 197 569 1998 32 2,389 9,335 2,290 9,068 99 266 1999 29 2,867 21,255 2,780 20,722 88 532 2000 41 2,863 19,678 2,786 19,188 77 489 2001 41 3,053 15,882 2,947 15,293 106 589 2002 26 2,733 20,830 2,636 20,143 97 686

23

Table A3. Pearson correlation coefficients between North Carolina’s scrap landings (bait_landed_pnds) and potential explanatory variables by gear type. nc_trips= number of trips ; nc_pnds=landings of Atlantic croaker; nc_unc_bait= unclassified finfish laded for bait or animal food; nc_unc_gen= unclassified finfish landed under the general category; nc_unc_food= unclassified finfish landed as food fish; nc_unc_total = nc_unc_bait+nc_unc_food+nc_unc_gen;

Gear Type Variable nc_pnds nc_trips nc_unc_food

nc_unc_bait

nc_unc_ gen

nc_unc_total

GILLNET bait_landed_pnds

-0.30 0.14 -0.25 -0.25 -0.05 -0.07

HAUL_S bait_landed_pnds

0.62 0.72 -0.12 0.67 0.23 0.67

OTTER_ TRAWL_F

bait_landed_pnds

-0.62 0.19 -0.19 0.41 0.40 0.46

POUND bait_landed_pnds

0.72 0.28 -0.27 0.68 0.23 0.68

Table A4. Summary of Stepwise regression models to examining the relationship of Scrap landings to potential explanatory variables for North Carolina. nc_trips= number of trips ; nc_pnds=landings of Atlantic croaker; nc_unc_bait= unclassified finfish laded for bait or animal food; nc_unc_gen= unclassified finfish landed under the general category. nc_unc_total = nc_unc_bait+nc_unc_food+nc_unc_gen.

Gear Type Dependent Step Var. Entered

Var.Removed

NumberIn

PartialR-Square

ModelR-square

Cp F Value Prob F

HAUL_S log_bait 1 nc_trips 1 0.476 0.476 4.633 13.650 0.002HAUL_S log_bait 2 nc_unc_bait 2 0.152 0.628 1.522 5.714 0.031HAUL_S log_bait 3 nc_pnds 3 0.029 0.657 2.552 1.091 0.315HAUL_S log_bait 4 nc_pnds 2 0.029 0.628 1.522 1.091 0.315OTTER_ TRAWL_F

log_bait 1 nc_pnds 1 0.442 0.442 3.633 11.882 0.004

OTTER_ TRAWL_F

log_bait 2 NC_UNC_GEN 2 0.167 0.609 0.644 5.999 0.028

POUND log_bait 1 nc_unc_total 1 0.457 0.457 8.173 12.621 0.003POUND log_bait 2 nc_trips 2 0.226 0.683 1.350 10.002 0.007POUND log_bait 3 nc_pnds 3 0.018 0.701 2.652 0.779 0.394POUND log_bait 4 nc_pnds 2 0.018 0.683 1.350 0.779 0.394GILLNET Inv_bait 1 nc_unc_bait 1 0.061 0.061 -0.944 0.976 0.339GILLNET Inv_bait 2 nc_unc_

bait 0 0.061 0.000 -2.159 0.976 0.339

24

Table A5. Parameter estimates for final stepwise regression model to estimate scrap landings of Atlantic croaker in North Carolina. nc_trips= number of trips ; nc_pnds=landings of Atlantic croaker; nc_unc_bait= unclassified finfish laded for bait or animal food; nc_unc_gen= unclassified finfish landed under the general category. nc_unc_total = nc_unc_bait+nc_unc_food+ nc_unc_gen. Gear Type Dep. Var Step Variable Estimate Std Err TypeII SS F-Value Prob. FHAUL_S log_bait 4 Intercept 11.99726 0.33352 584.30 1293.96 0.000HAUL_S log_bait 4 nc_trips 0.000865 0.000332 3.07 6.79 0.021HAUL_S log_bait 4 nc_unc_

bait 3.41E-07 1.43E-07 2.58 5.71 0.031

OTTER_ TRAWL_F

log_bait 2 Intercept 13.29162 0.253261 546.53 2754.35 0.000

OTTER_ TRAWL_F

log_bait 2 nc_pnds -1.6E-07 4.85E-08 2.27 11.46 0.004

OTTER_ TRAWL_F

log_bait 2 nc_unc_ gen

1.43E-06 5.83E-07 1.19 6.00 0.028

POUND log_bait 4 Intercept 9.429383 0.428459 518.13 484.34 0.000POUND log_bait 4 nc_unc_

total 2.35E-06 4.96E-07 23.95 22.39 0.000

POUND log_bait 4 nc_trips 0.000618 0.000196 10.70 10.00 0.007

25

Table A6. Unclassified Finfish landings for North Carolina (in MT) by gear. Also shown are the Landings of Atlantic croaker and the NCDMF estimated scrap landings for Atlantic croaker NC Unclassified Landings

Year Gear type Scrap Est Landings General Food Bait Total

1973 Gill net 166 0 0 0 0 1974 Gill net 133 0 0 0 0 1975 Gill net 61 0 0 0 0 1976 Gill net 59 0 0 0 0 1977 Gill net 376 0 0 0 0 1978 Gill net 547 0 0 0 0 1979 Gill net 901 0 0 1 1 1980 Gill net 1,712 17 0 0 17 1981 Gill net 603 109 0 0 109 1982 Gill net 562 23 0 0 23 1983 Gill net 406 3 0 0 3 1984 Gill net 1,127 8 0 0 8 1985 Gill net 1,085 10 0 0 10 1986 Gill net 0.5 1,526 35 0 0 35 1987 Gill net 0.5 1,200 57 0 0 57 1988 Gill net 0.5 1,198 1 0 0 1 1989 Gill net 0.5 642 14 0 0 14 1990 Gill net 13 396 37 0 0 37 1991 Gill net 6 385 11 0 0 11 1992 Gill net 7 465 19 0 0 19 1993 Gill net 0.5 384 168 9 0 177 1994 Gill net 0.5 665 2 3 0 6 1995 Gill net 0.5 941 5 3 4 11 1996 Gill net 0.5 1,944 5 2 0 7 1997 Gill net 0.5 1,315 4 3 2 8 1998 Gill net 3 2,616 2 3 1 6 1999 Gill net 0.5 1,817 2 2 2 6 2000 Gill net 8 1,769 2 3 0 5 2001 Gill net 0.5 2,436 1 2 0 4 2002 Gill net 2 1,921 5 1 0 6

26

Table A6 continued. NC Unclassified Landings

Year Gear type Scrap Est. Landings General Food Bait Total

1973 Haul Seine 1,114 0 0 902 902 1974 Haul Seine 1,616 0 0 960 960 1975 Haul Seine 2,941 0 0 1,406 1,406 1976 Haul Seine 2,000 0 0 1,170 1,170 1977 Haul Seine 3,510 0 0 1,336 1,336 1978 Haul Seine 3,165 0 0 2,150 2,150 1979 Haul Seine 4,190 0 0 2,524 2,524 1980 Haul Seine 3,720 31 0 2,870 2,901 1981 Haul Seine 2,487 42 0 1,940 1,982 1982 Haul Seine 2,175 19 0 1,954 1,973 1983 Haul Seine 1,951 21 0 2,065 2,085 1984 Haul Seine 1,490 59 0 2,054 2,113 1985 Haul Seine 1,117 29 0 1,164 1,193 1986 Haul Seine 298 1,399 41 0 1,424 1,465 1987 Haul Seine 838 708 10 0 1,393 1,403 1988 Haul Seine 600 1,203 12 0 1,078 1,090 1989 Haul Seine 897 1,461 29 0 1,122 1,151 1990 Haul Seine 880 1,771 7 0 1,557 1,565 1991 Haul Seine 495 899 30 0 856 886 1992 Haul Seine 66 426 9 0 195 204 1993 Haul Seine 232 202 15 1 108 124 1994 Haul Seine 633 55 9 0 45 54 1995 Haul Seine 823 79 2 0 65 67 1996 Haul Seine 81 208 4 0 38 42 1997 Haul Seine 142 38 3 0 49 51 1998 Haul Seine 70 17 6 0 16 22 1999 Haul Seine 111 14 4 0 11 15 2000 Haul Seine 200 32 6 0 6 12 2001 Haul Seine 110 49 7 0 0 7 2002 Haul Seine 53 15 6 0 2 8

27

Table A6 continued. NC Unclassified Landings

Year Gear type Scrap Est. Landings General Food Bait Total

1973 Trawl 580 0 0 2,463 2,463 1974 Trawl 851 0 0 3,868 3,868 1975 Trawl 1,414 0 0 1,396 1,396 1976 Trawl 3,732 0 0 1,367 1,367 1977 Trawl 4,426 0 0 1,563 1,563 1978 Trawl 4,943 0 0 1,082 1,082 1979 Trawl 3,662 0 0 4,757 4,757 1980 Trawl 2,510 18 0 2,262 2,280 1981 Trawl 894 11 0 1,584 1,594 1982 Trawl 832 13 0 1,095 1,108 1983 Trawl 370 5 0 1,419 1,424 1984 Trawl 1,031 44 0 1,765 1,809 1985 Trawl 995 21 1 1,606 1,627 1986 Trawl 132 995 24 1 1,507 1,532 1987 Trawl 281 724 213 1 1,003 1,217 1988 Trawl 598 866 150 0 2,172 2,322 1989 Trawl 433 622 101 1 659 761 1990 Trawl 156 155 7 0 40 48 1991 Trawl 365 137 59 0 22 80 1992 Trawl 578 342 56 0 0 56 1993 Trawl 234 859 144 2 2 148 1994 Trawl 259 1,351 242 11 90 343 1995 Trawl 325 1,688 214 16 21 250 1996 Trawl 302 2,126 248 4 15 266 1997 Trawl 197 3,252 55 6 45 106 1998 Trawl 99 2,289 121 5 0 126 1999 Trawl 88 2,777 3 4 22 29 2000 Trawl 77 2,786 11 5 0 16 2001 Trawl 106 2,947 9 3 0 12 2002 Trawl 97 2,611 13 3 0 16

28

Table A6 continued. NC Unclassified Landings Year Gear type Scrap Est. Landings General Food Bait Total

1973 Pound net 27 0 0 0 0 1974 Pound net 42 0 0 0 0 1975 Pound net 79 0 0 0 0 1976 Pound net 961 0 0 0 0 1977 Pound net 176 0 0 0 0 1978 Pound net 213 0 0 362 362 1979 Pound net 274 0 0 277 277 1980 Pound net 1,142 5 0 453 458 1981 Pound net 865 1 0 465 466 1982 Pound net 928 5 0 557 562 1983 Pound net 341 5 0 410 416 1984 Pound net 448 1 0 551 551 1985 Pound net 616 71 0 678 749 1986 Pound net 135 251 5 0 261 265 1987 Pound net 167 636 103 0 448 550 1988 Pound net 267 505 5 0 445 450 1989 Pound net 239 277 0 0 456 456 1990 Pound net 200 232 4 0 681 685 1991 Pound net 126 84 3 0 613 616 1992 Pound net 38 18 3 0 400 402 1993 Pound net 83 10 3 0 108 111 1994 Pound net 7 12 1 4 76 82 1995 Pound net 9 10 0 1 102 104 1996 Pound net 95 7 3 0 43 46 1997 Pound net 6 2 2 0 54 55 1998 Pound net 2 0 1 0 59 60 1999 Pound net 196 6 1 1 61 63 2000 Pound net 16 1 2 0 23 25 2001 Pound net 2 12 1 0 5 5 2002 Pound net 11 1 0 0 13 13

29

Table A7. Estimated Scrap landings for North Carolina Commercial Fisheries by gear. NC Estimates=NCDMF scrap estimates. reg=regression method. ratio=ratio method. land=based on landings. bait=based on unclassified finfish laded as bait or animal food. uncl= total unclassified fin fishes. All estimates are in MT.

NC Scrap Estimates Standard Error

Year Gear type Estimates reg -land reg-bait ratio-land ratio-bait ratio-uncl reg -land reg-bait ratio-land ratio-bait ratio-uncl

1973 Gillnet 2 1 1 0 0 0.05 0.02 0.02 0.00 0.001974 Gillnet 2 1 0 0 0 0.05 0.02 0.01 0.00 0.001975 Gillnet 2 1 0 0 0 0.05 0.02 0.01 0.00 0.001976 Gillnet 2 1 0 0 0 0.05 0.02 0.01 0.00 0.001977 Gillnet 2 1 1 0 0 0.06 0.02 0.04 0.00 0.001978 Gillnet 2 1 2 0 0 0.06 0.02 0.05 0.00 0.001979 Gillnet 2 1 3 0 0 0.10 0.02 0.08 0.00 0.001980 Gillnet 3 1 5 0 2 0.35 0.02 0.16 0.00 0.051981 Gillnet 2 1 2 0 11 0.07 0.02 0.06 0.00 0.291982 Gillnet 2 1 2 0 2 0.07 0.02 0.05 0.00 0.061983 Gillnet 2 1 1 0 0 0.08 0.02 0.04 0.00 0.011984 Gillnet 2 1 4 0 1 0.13 0.02 0.11 0.00 0.021985 Gillnet 2 1 3 0 1 0.11 0.02 0.10 0.00 0.031986 Gillnet 0.5 2 1 5 0 4 0.13 0.02 0.14 0.00 0.101987 Gillnet 0.5 2 1 3 0 6 0.10 0.02 0.11 0.00 0.161988 Gillnet 0.5 2 1 4 0 0 0.09 0.02 0.11 0.00 0.001989 Gillnet 0.5 2 1 2 0 1 0.07 0.02 0.06 0.00 0.041990 Gillnet 13.0 2 1 1 0 4 0.06 0.02 0.04 0.00 0.101991 Gillnet 6.0 2 1 1 0 1 0.06 0.02 0.04 0.00 0.031992 Gillnet 7.0 2 1 1 0 2 0.07 0.02 0.04 0.00 0.051993 Gillnet 0.5 2 1 1 0 18 0.06 0.02 0.04 0.00 0.491994 Gillnet 0.5 2 1 2 0 1 0.06 0.01 0.06 0.00 0.021995 Gillnet 0.5 2 1 3 0 1 0.10 0.02 0.09 0.00 0.031996 Gillnet 0.5 3 1 6 0 1 0.21 0.02 0.19 0.00 0.021997 Gillnet 0.5 2 1 4 0 1 0.16 0.02 0.12 0.00 0.021998 Gillnet 3.0 3 1 8 0 1 0.34 0.02 0.26 0.00 0.021999 Gillnet 0.5 3 1 6 0 1 0.22 0.02 0.17 0.00 0.022000 Gillnet 8.0 3 1 6 0 1 0.24 0.02 0.17 0.00 0.012001 Gillnet 0.5 3 1 7 0 0 0.26 0.02 0.23 0.00 0.012002 Gillnet 2.0 2 1 6 0 1 0.17 0.01 0.17 0.00 0.02

30

Table A7. Continued. NC Scrap Estimates Standard Error

Year Gear type Estimates reg-land reg-bait ratio-land ratio-bait ratio-uncl reg-land reg-bait ratio-land ratio-bait ratio-uncl

1973 Haul Seine 590 468 604 485 478 13.21 8.42 3.95 2.70 2.701974 Haul Seine 1,149 508 874 516 509 36.94 9.97 5.82 2.93 2.921975 Haul Seine 7,201 904 1,580 751 740 440.07 22.83 10.78 4.36 4.351976 Haul Seine 1,840 650 1,069 623 613 69.33 14.26 7.30 3.61 3.601977 Haul Seine 18,081 820 1,880 712 701 1491.06 21.18 12.79 4.12 4.111978 Haul Seine 9,628 2,495 1,698 1,148 1,131 566.08 88.62 11.21 6.45 6.431979 Haul Seine 49,601 4,239 2,232 1,339 1,319 6750.51 206.79 15.21 7.75 7.731980 Haul Seine 25,536 7,522 2,006 1,539 1,532 3170.59 488.30 13.15 8.58 8.651981 Haul Seine 3,746 1,904 1,340 1,039 1,046 186.47 64.58 8.93 5.89 6.001982 Haul Seine 2,154 1,778 1,158 1,036 1,031 83.96 56.24 7.41 5.63 5.671983 Haul Seine 1,729 2,230 1,047 1,103 1,097 61.52 76.33 7.05 6.31 6.361984 Haul Seine 963 2,276 808 1,106 1,121 29.42 88.30 5.20 6.07 6.231985 Haul Seine 561 630 597 619 625 12.22 12.99 3.85 3.40 3.471986 Haul Seine 298.0 801 888 745 756 766 21.41 22.05 4.93 4.25 4.361987 Haul Seine 838.0 336 835 375 736 730 5.72 20.42 2.46 4.09 4.111988 Haul Seine 600.0 646 579 647 576 574 16.12 12.15 4.22 3.20 3.231989 Haul Seine 897.0 875 599 777 594 601 24.51 12.67 5.36 3.48 3.561990 Haul Seine 880.0 1,437 1,156 965 843 834 48.16 31.56 6.23 4.64 4.651991 Haul Seine 495.0 439 431 483 457 466 8.51 7.48 3.18 2.56 2.651992 Haul Seine 66.0 252 190 230 105 108 3.49 1.94 1.55 0.60 0.631993 Haul Seine 232.0 190 169 108 58 65 2.01 1.49 0.71 0.32 0.371994 Haul Seine 633.0 161 156 29 24 28 1.59 1.40 0.21 0.14 0.171995 Haul Seine 823.0 168 162 43 35 35 1.68 1.46 0.29 0.21 0.211996 Haul Seine 81.0 193 156 112 20 22 2.17 1.32 0.76 0.12 0.131997 Haul Seine 142.0 159 158 20 26 27 1.45 1.34 0.13 0.15 0.151998 Haul Seine 70.0 156 152 9 9 12 1.41 1.26 0.06 0.05 0.071999 Haul Seine 111.0 154 151 7 6 8 1.32 1.19 0.05 0.03 0.042000 Haul Seine 200.0 159 151 17 3 6 1.52 1.28 0.12 0.02 0.042001 Haul Seine 110.0 162 150 27 0 4 1.53 1.22 0.18 0.00 0.022002 Haul Seine 53.0 155 149 8 1 4 1.39 1.21 0.05 0.00 0.02

31

Table A7 Continued. NC Scrap Estimates Standard Error

Year Gear type Estimates reg-land reg-bait ratio-land ratio-bait ratio-uncl reg-land reg-bait ratio-land ratio-bait ratio-uncl

1973 Trawl 327 658 276 733 671 3.15 18.60 2.60 6.88 5.881974 Trawl 293 1,566 403 1,143 1,047 3.13 66.00 3.82 10.81 9.231975 Trawl 241 359 680 419 384 3.06 6.09 6.13 3.77 3.221976 Trawl 113 366 1,819 416 380 2.74 6.39 16.94 3.86 3.301977 Trawl 81 380 2,082 459 421 2.28 7.13 19.64 4.32 3.691978 Trawl 71 302 2,363 323 296 2.04 4.28 22.13 3.01 2.571979 Trawl 112 3,175 1,771 1,437 1,314 2.61 185.86 16.12 13.03 11.131980 Trawl 169 625 1,213 683 629 3.36 18.54 11.68 6.55 5.641981 Trawl 290 392 425 470 433 3.10 7.23 4.03 4.44 3.821982 Trawl 290 293 386 318 295 3.10 4.33 3.83 3.14 2.711983 Trawl 350 351 174 417 383 2.99 5.94 1.63 3.89 3.331984 Trawl 279 446 495 530 497 3.26 9.84 4.56 4.86 4.261985 Trawl 280 401 473 477 443 3.23 8.02 4.51 4.53 3.931986 Trawl 132.0 279 376 473 448 417 3.13 6.95 4.40 4.15 3.601987 Trawl 281.0 313 293 347 301 334 3.21 4.09 3.33 2.87 2.971988 Trawl 598.0 293 549 411 644 630 3.17 13.73 3.89 6.08 5.551989 Trawl 433.0 320 241 295 196 207 3.06 2.62 2.72 1.79 1.771990 Trawl 156.0 384 180 74 12 13 2.93 1.08 0.68 0.11 0.111991 Trawl 365.0 384 177 64 6 22 2.99 1.07 0.62 0.06 0.191992 Trawl 578.0 362 178 164 0 15 3.12 1.05 1.55 0.00 0.131993 Trawl 234.0 294 177 409 1 40 3.13 1.03 3.83 0.01 0.351994 Trawl 259.0 244 184 645 27 93 2.87 1.13 5.76 0.24 0.781995 Trawl 325.0 217 179 807 6 68 2.96 1.04 7.37 0.06 0.581996 Trawl 302.0 186 178 1,012 4 72 3.16 1.10 9.83 0.04 0.651997 Trawl 197.0 126 181 1,560 13 29 2.50 1.10 14.54 0.13 0.251998 Trawl 99.0 176 177 1,098 0 34 2.91 1.02 10.12 0.00 0.301999 Trawl 88.0 148 179 1,328 6 8 2.85 1.07 12.39 0.06 0.072000 Trawl 77.0 149 178 1,337 0 4 2.87 1.04 12.60 0.00 0.042001 Trawl 106.0 133 175 1,384 0 3 2.50 0.98 12.61 0.00 0.032002 Trawl 97.0 154 176 1,231 0 4 2.87 1.05 12.02 0.00 0.04

32

Table A7. Continued NC Scrap Estimates Standard Error

Year Gear type Estimates reg-land reg-bait ratio-land ratio -bait ratio-uncl reg-land reg-bait ratio -land ratio-bait ratio-uncl

1973 Pound net 25 13 14 0 0 0.38 0.20 0.10 0.00 0.001974 Pound net 29 13 23 0 0 0.49 0.21 0.17 0.00 0.001975 Pound net 37 14 43 0 0 0.72 0.21 0.30 0.00 0.001976 Pound net 23,130 13 509 0 0 2766.23 0.19 3.64 0.00 0.001977 Pound net 64 13 93 0 0 1.72 0.20 0.66 0.00 0.001978 Pound net 85 112 113 159 152 2.78 5.02 0.82 0.74 0.761979 Pound net 128 66 146 122 116 4.22 1.97 1.04 0.56 0.571980 Pound net 138,724 179 593 197 189 28782.04 8.26 4.45 0.94 0.971981 Pound net 17,001 216 460 205 195 4282.51 12.37 3.32 0.95 0.971982 Pound net 23,748 378 492 245 235 4642.46 22.37 3.55 1.13 1.171983 Pound net 214 156 181 181 174 9.93 7.11 1.35 0.86 0.891984 Pound net 477 394 240 244 232 23.95 19.51 1.75 1.15 1.171985 Pound net 1,574 836 325 297 312 110.89 50.84 2.47 1.45 1.631986 Pound net 135.0 102 56 132 114 110 3.20 1.61 0.95 0.53 0.551987 Pound net 167.0 1,780 186 339 198 231 121.10 7.25 2.46 0.92 1.161988 Pound net 267.0 659 181 267 195 188 39.04 7.69 1.96 0.92 0.951989 Pound net 239.0 127 194 146 200 191 4.58 9.10 1.07 0.94 0.961990 Pound net 200.0 98 993 124 301 288 3.56 105.82 0.90 1.41 1.451991 Pound net 126.0 35 593 44 267 255 0.77 59.21 0.33 1.29 1.331992 Pound net 38.0 24 143 10 177 169 0.36 6.12 0.07 0.82 0.851993 Pound net 83.0 23 24 5 48 46 0.33 0.51 0.04 0.23 0.241994 Pound net 7.0 23 20 6 33 34 0.35 0.42 0.05 0.16 0.181995 Pound net 9.0 23 24 5 45 44 0.34 0.51 0.04 0.22 0.231996 Pound net 95.0 22 17 4 19 19 0.30 0.28 0.03 0.09 0.091997 Pound net 6.0 22 18 1 24 23 0.30 0.31 0.01 0.11 0.121998 Pound net 2.0 22 19 0 26 25 0.30 0.34 0.00 0.12 0.131999 Pound net 196.0 22 18 3 27 26 0.30 0.33 0.02 0.12 0.132000 Pound net 16.0 22 15 1 10 11 0.29 0.24 0.00 0.05 0.052001 Pound net 2.0 23 14 6 2 2 0.35 0.22 0.05 0.01 0.012002 Pound net 11.0 21 14 1 6 5 0.30 0.23 0.00 0.03 0.03

33

Table A8. Unclassified Finfish landings for Virginia (in MT) by gear. Also shown are the Landings of Atlantic croaker and the NCDMF scrap cpue (per trip in Kg) and where available the number of commercial trips for Atlantic croaker.

Year Gear type Landings General Spawn Food Bait Total Trips NC bait/trip

(kgs)1973 Gill net 115 0 0 7 343 350 01974 Gill net 94 0 0 4 14 18 01975 Gill net 303 0 0 1 26 27 01976 Gill net 701 0 0 1 5 6 01977 Gill net 936 0 0 4 25 29 01978 Gill net 566 0 0 7 102 109 01979 Gill net 167 0 0 6 53 59 01980 Gill net 54 0 0 2 20 22 01981 Gill net 18 0 0 1 10 10 01982 Gill net 11 0 0 3 46 49 01983 Gill net 13 0 0 2 29 31 01984 Gill net 58 0 0 1 64 65 01985 Gill net 195 0 0 2 67 68 01986 Gill net 206 0 0 1 47 48 01987 Gill net 175 0 0 0 96 96 01988 Gill net 264 0 0 1 42 43 01989 Gill net 164 0 0 22 32 54 0.51990 Gill net 47 26 0 2 53 81 0.51991 Gill net 55 5 0 7 53 65 0.51992 Gill net 332 4 0 4 57 65 101993 Gill net 1342 0 0 2 62 64 2166 0.51994 Gill net 1441 1 0 1 153 155 1677 11995 Gill net 1284 0 0 0 237 237 1598 0.51996 Gill net 1736 2 0 1 217 220 1390 0.51997 Gill net 2323 6 0 1 175 182 1813 0.51998 Gill net 2177 13 0 1 296 310 1943 11999 Gill net 1578 37 0 0 121 158 1541 0.52000 Gill net 2710 58 0 0 160 218 1770 12001 Gill net 2356 0 0 1 84 84 1528 0.52002 Gill net 2112 4 0 0 150 153 1598 0.5

34

Table A8. continued.

Year Gear type Landings General Spawn Food Bait Total Trips NC bait/trip

(kgs)1973 Haul Seine 201 0 0 0 31 32 01974 Haul Seine 45 0 0 0 9 10 01975 Haul Seine 179 0 0 0 85 85 01976 Haul Seine 284 0 0 0 14 15 01977 Haul Seine 422 0 0 5 14 20 01978 Haul Seine 302 0 0 2 46 48 01979 Haul Seine 183 0 0 0 299 299 01980 Haul Seine 21 0 0 0 67 67 01981 Haul Seine 29 0 0 0 77 77 01982 Haul Seine 0 0 0 0 36 37 01983 Haul Seine 5 0 0 0 54 54 01984 Haul Seine 81 0 0 0 84 84 01985 Haul Seine 504 0 0 1 129 130 01986 Haul Seine 598 0 0 0 137 137 4851987 Haul Seine 807 0 0 0 35 36 3351988 Haul Seine 315 0 0 0 10 10 3771989 Haul Seine 126 0 0 0 88 88 4021990 Haul Seine 5 0 0 0 43 43 6641991 Haul Seine 7 0 0 0 160 160 2781992 Haul Seine 205 0 0 0 290 290 341993 Haul Seine 384 0 0 0 274 274 399 1671994 Haul Seine 484 0 0 0 330 330 378 5911995 Haul Seine 581 0 0 0 257 257 324 4421996 Haul Seine 695 5 0 0 238 243 358 1831997 Haul Seine 1438 0 0 0 138 138 490 1801998 Haul Seine 1060 0 0 1 418 419 522 1451999 Haul Seine 1287 0 0 0 319 319 512 1472000 Haul Seine 955 0 0 0 215 215 397 2152001 Haul Seine 1006 0 0 0 117 117 402 10332002 Haul Seine 1237 1 0 0 168 169 370 301

35

Table A8. Continued.

Year Gear type Landings General Spawn Food Bait Total Trips NC bait/trip

(kgs)1973 Trawl 140 0 0 2 0 2 01974 Trawl 218 0 0 2 5 7 01975 Trawl 606 0 0 4 1 5 01976 Trawl 420 0 0 12 0 12 01977 Trawl 295 0 0 9 2 11 01978 Trawl 379 0 0 7 0 7 01979 Trawl 99 0 0 1 0 1 01980 Trawl 29 0 0 1 3 5 01981 Trawl 11 0 0 1 0 1 01982 Trawl 11 0 0 0 20 21 01983 Trawl 7 0 0 0 3 3 01984 Trawl 48 0 0 1 2 3 01985 Trawl 62 0 0 2 0 2 01986 Trawl 36 0 0 0 0 0 8181987 Trawl 34 0 0 0 5 6 11811988 Trawl 10 0 0 0 1 1 13631989 Trawl 26 0 0 0 20 20 14671990 Trawl 0 0 0 0 0 1 5421991 Trawl 3 0 0 1 1 3 5971992 Trawl 10 0 0 0 1 1 14551993 Trawl 62 0 0 0 0 0 16371994 Trawl 62 0 0 0 0 0 20871995 Trawl 112 1 0 18 0 19 14751996 Trawl 193 776 0 15 0 792 13051997 Trawl 425 734 0 2 0 736 5691998 Trawl 311 636 0 0 0 636 2661999 Trawl 612 289 0 9 0 298 5322000 Trawl 515 27 0 0 0 27 4892001 Trawl 480 0 0 3 0 3 5892002 Trawl 439 1 0 0 0 1 686

36

Table A8. Continued.

Year Gear type Landings General Spawn Food Bait Total Trips NC bait/trip

(kgs)1973 Pound net 160 0 0 17 4261 4278 01974 Pound net 322 0 0 77 1943 2020 01975 Pound net 1053 0 0 6 2980 2986 01976 Pound net 1262 0 0 5 3715 3721 01977 Pound net 2236 0 0 8 4383 4391 01978 Pound net 2424 0 0 6 6618 6624 01979 Pound net 516 0 0 17 5208 5224 01980 Pound net 218 0 0 5 1806 1812 01981 Pound net 137 0 0 2 915 917 01982 Pound net 32 0 0 2 1189 1191 01983 Pound net 43 0 0 1 1810 1811 01984 Pound net 183 0 0 0 894 895 01985 Pound net 225 0 0 1 766 767 01986 Pound net 233 0 0 1 396 397 2501987 Pound net 218 0 0 2 465 467 1351988 Pound net 204 0 0 2 214 216 1241989 Pound net 112 0 0 4 909 913 2541990 Pound net 37 0 0 3 507 510 2911991 Pound net 9 0 0 3 685 689 1951992 Pound net 60 0 0 5 848 852 601993 Pound net 595 0 0 31 2405 2436 1580 351994 Pound net 615 0 0 0 2769 2769 1607 571995 Pound net 1178 0 0 0 3351 3352 2228 81996 Pound net 1642 9 0 0 2835 2844 2113 171997 Pound net 1592 12 0 0 3310 3322 2502 111998 Pound net 1852 5 0 3 1422 1430 3234 251999 Pound net 2324 6 0 0 1539 1545 2781 1322000 Pound net 1638 4 0 6 1909 1920 2614 1262001 Pound net 1997 0 0 0 1347 1348 2236 352002 Pound net 1617 5 0 3 1246 1255 2238 21

37

Table A9. Estimated Scrap landings from Virginia commercial Fisheries by gear. reg=regression method. ratio=ratio method. land=based on landings. bait=based on unclassified finfish laded as bait or animal food. uncl= total unclassified fin fishes. step=based on stepwise regression. va-cpue/trip= using NCDMF scrap cpue/trip * number of trips from trip ticket database. Note 2nd column has different number of trips (see text) . All estimates in MT. Estimate Std Error

Year gear type reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step va-cpue/trip reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step1973 Gill net 2 22 0 37 0.05 6.02 0.01 0.00 0.98 1974 Gill net 2 1 0 2 0.04 0.02 0.01 0.00 0.05 1975 Gill net 2 1 1 3 0.05 0.02 0.03 0.00 0.07 1976 Gill net 2 1 2 1 0.09 0.02 0.07 0.00 0.02 1977 Gill net 2 1 3 3 0.10 0.03 0.09 0.00 0.08 1978 Gill net 2 1 2 11 0.06 0.07 0.05 0.00 0.29 1979 Gill net 2 1 0 6 0.05 0.04 0.02 0.00 0.16 1980 Gill net 2 1 0 2 0.05 0.03 0.01 0.00 0.06 1981 Gill net 2 1 0 1 0.05 0.02 0.00 0.00 0.03 1982 Gill net 2 1 0 5 0.05 0.04 0.00 0.00 0.14 1983 Gill net 2 1 0 3 0.05 0.03 0.00 0.00 0.08 1984 Gill net 2 1 0 7 0.05 0.05 0.01 0.00 0.18 1985 Gill net 2 1 1 7 0.05 0.05 0.02 0.00 0.19 1986 Gill net 2 1 1 5 0.05 0.03 0.02 0.00 0.13 1987 Gill net 2 1 0 10 0.05 0.06 0.02 0.00 0.26 1988 Gill net 2 1 1 4 0.05 0.03 0.02 0.00 0.12 1989 Gill net 2 1 0 5 0.05 0.03 0.02 0.00 0.15 1990 Gill net 2 1 0 8 0.05 0.04 0.00 0.00 0.22 1991 Gill net 2 1 0 7 0.05 0.04 0.01 0.00 0.18 1992 Gill net 2 1 1 7 0.06 0.04 0.03 0.00 0.18 1993 Gill net 2 1 4 7 1 0.13 0.04 0.13 0.00 0.18 1994 Gill net 2 2 4 15 2 0.09 0.10 0.13 0.00 0.42 1995 Gill net 2 5 4 25 1 0.13 0.92 0.12 0.00 0.66 1996 Gill net 3 4 6 24 1 0.18 0.47 0.17 0.00 0.61 1997 Gill net 3 3 7 19 1 0.47 0.40 0.22 0.00 0.49 1998 Gill net 3 7 6 31 2 0.22 2.01 0.21 0.00 0.88 1999 Gill net 3 2 5 17 1 0.17 0.11 0.15 0.00 0.44 2000 Gill net 5 3 8 23 2 0.66 0.27 0.26 0.00 0.61 2001 Gill net 3 1 7 9 1 0.24 0.05 0.22 0.00 0.23 2002 Gill net 3 2 6 15 1 0.20 0.13 0.19 0.00 0.40

38

Table A9. Continued. Estimate Std Error

Year gear type reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step va-cpue/trip reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step1973 Haul Seine 194 156 109 17 17 2.10 1.28 0.71 0.09 0.09 1974 Haul Seine 162 152 24 5 5 1.51 1.25 0.16 0.03 0.03 1975 Haul Seine 187 165 96 45 45 2.04 1.49 0.66 0.26 0.26 1976 Haul Seine 210 151 152 8 8 2.56 1.24 1.04 0.04 0.04 1977 Haul Seine 247 151 226 8 10 3.55 1.27 1.54 0.04 0.06 1978 Haul Seine 214 157 162 24 25 2.61 1.32 1.07 0.14 0.14 1979 Haul Seine 185 212 97 159 157 2.03 2.42 0.66 0.92 0.92 1980 Haul Seine 156 162 11 36 35 1.39 1.40 0.07 0.20 0.20 1981 Haul Seine 158 164 16 41 40 1.43 1.45 0.10 0.23 0.23

1982 Haul Seine 150 153 0 19 19 1.24 1.22 0.00 0.11 0.11 1983 Haul Seine 153 159 3 29 29 1.34 1.36 0.02 0.17 0.17 1984 Haul Seine 168 166 44 45 45 1.60 1.46 0.28 0.25 0.25 1985 Haul Seine 268 172 269 69 68 3.78 1.55 1.74 0.38 0.38 1986 Haul Seine 299 173 319 73 72 4.69 1.61 2.11 0.41 0.41 1987 Haul Seine 378 152 428 19 19 6.92 1.24 2.80 0.10 0.10 1988 Haul Seine 218 151 169 5 5 2.72 1.21 1.11 0.03 0.03 1989 Haul Seine 173 164 67 47 46 1.78 1.48 0.46 0.27 0.27 1990 Haul Seine 156 159 3 23 23 1.35 1.33 0.02 0.13 0.13 1991 Haul Seine 153 180 4 85 84 1.32 1.72 0.03 0.48 0.48 1992 Haul Seine 194 214 111 155 153 2.17 2.41 0.75 0.89 0.89 1993 Haul Seine 235 206 206 146 144 149 67 2.96 2.17 1.34 0.81 0.81 2.851994 Haul Seine 265 221 258 175 173 146 223 4.10 2.73 1.82 1.05 1.05 2.641995 Haul Seine 304 205 313 138 136 136 143 5.03 2.31 2.15 0.81 0.80 2.431996 Haul Seine 345 199 373 127 128 143 66 6.02 2.12 2.52 0.73 0.75 2.531997 Haul Seine 883 176 774 74 73 144 88 26.43 1.67 5.08 0.41 0.41 2.711998 Haul Seine 546 250 571 224 221 189 76 12.62 3.21 3.86 1.29 1.29 4.091999 Haul Seine 711 219 692 170 168 178 75 17.29 2.43 4.52 0.95 0.95 3.822000 Haul Seine 485 195 514 115 114 145 85 10.69 2.13 3.63 0.69 0.69 2.742001 Haul Seine 517 173 544 63 62 128 415 11.27 1.63 3.67 0.36 0.36 2.142002 Haul Seine 680 183 664 90 89 127 111 16.77 1.81 4.55 0.52 0.52 2.08

39

Table A9. Continued Estimate Std Error

Year gear type reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step va-cpue/trip reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step1973 Trawl 387 177 67 0 1 267 3.03 1.04 0.63 0.00 0.00 2.291974 Trawl 374 176 103 2 2 261 3.04 1.04 0.98 0.01 0.02 2.381975 Trawl 326 178 291 0 1 228 3.06 1.01 2.63 0.00 0.01 2.391976 Trawl 356 179 205 0 3 240 3.23 1.07 1.91 0.00 0.03 2.361977 Trawl 360 175 139 1 3 251 2.99 1.02 1.31 0.01 0.03 2.401978 Trawl 354 177 181 0 2 244 3.04 1.02 1.70 0.00 0.02 2.311979 Trawl 398 178 48 0 0 266 3.01 1.03 0.44 0.00 0.00 2.301980 Trawl 410 179 14 1 1 276 3.22 1.12 0.13 0.01 0.01 2.271981 Trawl 406 176 5 0 0 278 2.94 1.03 0.05 0.00 0.00 2.291982 Trawl 400 176 5 6 6 273 2.96 1.07 0.05 0.06 0.05 2.131983 Trawl 403 175 3 1 1 277 2.88 1.01 0.03 0.01 0.01 2.271984 Trawl 404 178 23 1 1 276 3.04 1.05 0.21 0.01 0.01 2.271985 Trawl 399 177 30 0 0 269 3.04 1.05 0.28 0.00 0.00 2.181986 Trawl 402 176 17 0 0 271 2.95 1.02 0.16 0.00 0.00 2.221987 Trawl 406 178 16 2 2 270 3.04 1.07 0.16 0.02 0.01 2.181988 Trawl 406 176 5 0 0 277 2.98 1.04 0.04 0.00 0.00 2.261989 Trawl 403 178 13 6 5 272 2.90 1.05 0.12 0.05 0.05 2.281990 Trawl 408 176 0 0 0 283 2.87 1.01 0.00 0.00 0.00 2.301991 Trawl 404 175 2 0 1 277 2.93 1.03 0.02 0.00 0.01 2.231992 Trawl 411 178 5 0 0 278 3.00 1.05 0.05 0.00 0.00 2.321993 Trawl 399 177 30 0 0 271 2.97 1.03 0.28 0.00 0.00 2.191994 Trawl 399 177 30 0 0 270 2.79 0.97 0.26 0.00 0.00 2.401995 Trawl 392 177 54 0 5 270 2.91 1.00 0.49 0.00 0.04 2.421996 Trawl 380 177 92 0 216 6,681 3.15 1.07 0.89 0.00 1.95 433.301997 Trawl 349 177 204 0 202 5,250 3.03 1.02 1.90 0.00 1.75 352.611998 Trawl 365 177 149 0 174 3,403 3.05 1.02 1.37 0.00 1.49 175.251999 Trawl 324 177 292 0 81 672 3.15 1.03 2.73 0.00 0.71 15.692000 Trawl 338 178 247 0 7 260 3.16 1.05 2.33 0.00 0.06 3.102001 Trawl 334 175 225 0 1 239 2.92 0.98 2.06 0.00 0.01 2.412002 Trawl 342 1.8E+02 207 0 0 241 3.13 1.05 2.02 0.00 0.00 2.40

40

Table A9 continued. Estimate Std Error

Year gear type reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step va-cpue/trip reg -land reg-bait ratio-land ratio-bait ratio-uncl reg-step1973 Pound net 56 1.4E+16 84 1,858 1,774 1.45 8.87E+15 0.62 8.79 9.03 1974 Pound net 190 3.2E+07 172 859 850 8.60 1.48E+07 1.29 4.16 4.43 1975 Pound net 104,741 3.2E+11 572 1,333 1,272 2.06E+04 1.68E+11 4.05 6.10 6.25 1976 Pound net 311,105 2.3E+13 668 1,633 1,556 5.95E+04 1.58E+13 4.78 7.49 7.67 1977 Pound net 1.6E+10 1.1E+17 1,186 1,929 1,838 1.13E+10 9.88E+16 8.43 8.79 9.00 1978 Pound net 1.8E+12 6.0E+28 1,291 2,920 2,781 1.71E+12 6.01E+28 9.37 13.61 13.93 1979 Pound net 742 7.8E+18 275 2,301 2,197 41.38 4.08E+18 1.96 10.55 10.82 1980 Pound net 81 4.0E+06 113 784 748 2.49 1.12E+06 0.85 3.74 3.84 1981 Pound net 51 5,203 73 403 384 1.25 890.50 0.52 1.87 1.91 1982 Pound net 26 35,607 17 523 499 0.40 6.90E+03 0.12 2.42 2.48 1983 Pound net 28 7.4E+06 23 798 759 0.49 2.17E+06 0.17 3.79 3.88 1984 Pound net 71 4,203 98 396 377 1.93 315.79 0.72 1.86 1.90 1985 Pound net 90 1,522 118 336 320 2.67 104.91 0.90 1.64 1.67 1986 Pound net 90 126 122 173 165 2.70 4.86 0.88 0.80 0.82 1987 Pound net 86 207 116 205 196 2.36 8.30 0.84 0.96 0.98 1988 Pound net 77 44 108 94 90 2.17 1.17 0.79 0.44 0.46 1989 Pound net 43 4,190 59 399 381 0.91 466.96 0.43 1.86 1.92 1990 Pound net 27 296 20 224 214 0.46 20.08 0.14 1.05 1.08 1991 Pound net 22 992 5 298 285 0.34 119.22 0.03 1.44 1.48 1992 Pound net 31 2,987 32 374 358 0.56 307.31 0.23 1.75 1.80 1993 Pound net 1,398 7.5E+08 315 1,057 1,019 7.1E+08 55 105.74 3.49E+08 2.34 5.04 5.22 2.29E+081994 Pound net 1,842 4.5E+10 325 1,216 1,157 4.5E+09 92 176.60 1.96E+10 2.46 5.89 6.03 1.40E+091995 Pound net 250,996 2.2E+12 631 1,484 1,412 1.7E+12 18 3.77E+04 7.99E+11 4.77 7.22 7.38 6.80E+111996 Pound net 1.8E+07 3.0E+10 874 1,250 1,193 9.5E+10 36 6.99E+06 1.53E+10 6.16 5.66 5.81 3.59E+101997 Pound net 9.1E+06 1.1E+12 847 1,459 1,394 9.6E+11 28 2.37E+06 4.71E+11 6.27 6.94 7.12 3.21E+111998 Pound net 3.1E+08 241,188 988 628 601 1.6E+06 81 2.06E+08 6.52E+04 7.26 2.96 3.05 3.37E+051999 Pound net 1.8E+10 473,857 1,231 677 647 2.9E+06 367 7.51E+09 9.33E+04 8.80 3.10 3.18 6.27E+052000 Pound net 2.9E+07 1.4E+07 871 841 805 4.9E+07 329 1.85E+07 7.63E+06 6.11 3.79 3.90 1.67E+072001 Pound net 3.1E+09 162,623 1,061 594 565 231,969 78 2.43E+09 4.89E+04 7.99 2.87 2.93 3.82E+042002 Pound net 1.9E+07 56,157 851 546 523 177,427 47 7.26E+06 9.31E+03 6.54 2.68 2.76 3.63E+04

41

APPENDIX B: Estimating Virginia’s Scrap Landings: Using Virginia field sampling data Scrap or bait landings in Virginia comprise small-size croaker harvested primarily from small-mesh haul seine and pound net fisheries. However, gill net, out-of-state trawl fisheries (Virginia has prohibited trawling in its state waters, since 1989), and other non-directed gear types (e. g. pots, dredges) that harvest a minor amount of Atlantic croaker also contribute to the scrap component of croaker landings. The Virginia Marine Resource Commission (VMRC) has collected samples size (length, weight) data from its commercial fisheries since 1989, by gear and market category. Though not differentiated, these data include information on both the marketable and scrap component. Classification of Virginia Atlantic croaker landings by market grade (unclassified, small, medium and large) was initiated in 1989. Atlantic croaker landings from all gear types combined show a decline in small-grade Atlantic croaker and corresponding general increase in large-grade Atlantic croaker over time, 1989-2003 (Table B1). This decline in small-grade Atlantic croaker is tied to the general population increase in numbers across ages, evident since the mid-1990s. Additionally, of the major gears responsible for landing small-grade Atlantic croaker (haul seine and pound net), the number of active pound nets has declined by 45%, since 1994. Haul seine trips during the last 10 years have fluctuated but are similar. Based on mesh size characteristics, haul seine and pound net gears in Virginia, are responsible for the majority of small (< 9 inches) Atlantic croaker harvested. By extension, these two gear types can be expected to contribute the most to the scrap (or bait) component of Atlantic croaker landings, and landings from each gear type show a decline in the proportion of small-grade Atlantic croaker during the 1989-2003 period (Tables B2 and B3). The market grade classification system of landings allowed for a proportional expansion of Atlantic croaker lengths (converted to weights) vs. Atlantic croaker landings to estimate total scrap (pounds), for the 1989 – 2002 period. It was decided that any Atlantic croaker less than 9 inches would be considered as potential scrap, and that ½ of the Atlantic croaker within a 9 to 9.99-inch length interval would also contribute to the scrap component (crab bait or other uses). Lengths of Atlantic croaker that satisfied their inclusion in the scrap category were converted to weights, using a length weight relationship. Total scrap landings were determined by using the proportion of scrap to total weight by year, gear and market grade were used to apportion the Atlantic croaker landings by year gear and market category (Table B4). Scrap estimates for the period 1973 to 1988 were estimated using an average ratio of estimated scrap to landings by gear from 1989-1993 and applying it to total landings by gear. These data was also used to develop estimates in numbers and a size distribution in 20mm length classes (Table B5). Using the field sample of lengths from the Virginia harvest to estimate Virginia scrap is preferable to using data from North Carolina because there are distinct regional differences among the gear, area, and seasonal contributions to the Atlantic croaker landings and scrap. For example, the majority of the scrap in North Carolina stems from

42

ocean trawl fisheries in coastal waters during late fall through winter, whereas the Virginia scrap primarily represents harvest from inside waters by pound net and haul seine fisheries during spring through late summer. The other approach to estimating the Virginia scrap component of Atlantic croaker landings used the ratio of North Carolina’s scrap to NMFS unclassified (all species) bait category to apportion Virginia’s NMFS unclassified (all species) bait category. However, using this method assumes that the relationship estimated for North Carolina is also appropriate for Virginia, which may not be a suitable assumption given the regional gear differences between the states. Potential limitations in using Virginia sample data to determine scrap include a potential for not sampling some small (< 7 inches) Atlantic croaker that are immediately set aside at the dock by the harvester for bait use. Also, the choice of assigning ½ of Atlantic croaker in the 9-inch interval (average weight ranged from 0.35 to of 0.42 pounds during 1989 – 2002) to the scrap component was initially based on the VMRC understanding of the marketing factors that change over time and within a season. The VMRC contacted long-time, high-volume seafood buyers (one on the western and one on the eastern shore) that wholesale Atlantic croaker from pound nets. The buyers indicated that Atlantic croaker less than 9 inches could generally be considered as scrap. However, both buyers and a middle peninsula buyer indicated that some small-size Atlantic croaker (< 9 inches) was sold for food during years of low Atlantic croaker abundance. The buyers generally agreed that ½ of Atlantic croaker within the 9-inch interval are sold as food fish, with a greater amount of this size category in the bait in recent years and less in earlier years. Recommendation: The technical committee endorsed using Atlantic croaker length data, collected by the VMRC, as the best method for estimating the scrap component of Atlantic croaker landings in Virginia.

43

Table B1. Pounds landed and percentages of total landings (pounds) of Atlantic croaker, by market category and year, for all gear types combined.

Jumbo Large Medium Small Unclassified

Year Pounds % Pounds % Pounds % Pounds % Pounds % Total

1989 27,266 2.87% 280,111 29.50% 252,854 26.63% 389,418 41.01% 949,649

1990 135 0.07% 48,706 24.19% 20,556 10.21% 131,956 65.53% 201,353

1991 3,031 1.85% 12,319 7.51% 40,243 24.52% 108,533 66.13% 164,126

1992 39,190 2.93% 144,618 10.80% 522,894 39.04% 632,651 47.24% 1,339,353

1993 236,537 4.50% 1,012,462 19.25% 875,599 16.65% 3,135,589 59.61% 5,260,187

1994 219,803 3.82% 1,078,449 18.72% 851,764 14.79% 3,609,959 62.67% 5,759,975

1995 399,929 5.75% 1,126,052 16.20% 934,840 13.45% 4,488,818 64.59% 6,949,639

1996 760,685 8.08% 1,375,274 14.62% 786,297 8.36% 6,487,648 68.94% 9,409,904

1997 1,562,283 12.15% 2,153,049 16.74% 1,236,503 9.61% 7,908,809 61.50% 12,860,644

1998 2,849,047 23.59% 2,315,277 19.17% 991,497 8.21% 5,920,847 49.03% 12,076,668

1999 3,036,692 23.60% 2,691,293 20.91% 1,175,712 9.14% 5,964,722 46.35% 12,868,419

2000 3,379,066 25.88% 2,086,243 15.98% 1,467,594 11.24% 6,121,281 46.89% 13,054,184

2001 3,299,005 25.32% 2,423,227 18.60% 1,072,696 8.23% 6,232,927 47.84% 13,027,855

2002 76,790 0.63% 3,149,279 25.87% 2,591,993 21.29% 568,846 4.67% 5,786,789 47.54% 12,173,697

2003 32,118 0.29% 3,208,629 29.34% 2,957,566 27.04% 492,252 4.50% 4,246,076 38.82% 10,936,641

44

Table B2 Virginia landings of Atlantic croaker from pound net, by market grade and year (1973-2003). Included are percentages of total pound net landings, by market grade.

Jumbo % Large % Medium % Small % Unclassified %Year pounds Total pounds Total pounds Total pounds Total pounds Total Total1973 349,343 100% 349,3431974 704,081 100% 704,0811975 2,281,840 100% 2,281,8401976 2,560,877 100% 2,560,8771977 4,024,832 100% 4,024,8321978 4,990,645 100% 4,990,6451979 1,081,930 100% 1,081,9301980 480,617 100% 480,6171981 300,560 100% 300,5601982 70,477 100% 70,4771983 94,476 100% 94,4761984 365,394 100% 365,3941985 486,751 100% 486,7511986 431,690 100% 431,6901987 402,005 100% 402,0051988 436,950 100% 436,9501989 21,506 9% 80,764 33% 55,564 22% 89,632 36% 247,4661990 10 0.01% 2,078 3% 2,220 3% 77,546 95% 81,8541991 379 2% 34 0.2% 576 3% 18,477 95% 19,4661992 4,985 4% 10,179 8% 23,981 18% 93,666 71% 132,8111993 89,721 7% 144,555 11% 190,638 15% 886,752 68% 1,311,6661994 125,171 9% 218,734 16% 222,516 16% 787,194 58% 1,353,6151995 284,552 11% 315,033 12% 296,226 11% 1,699,673 65% 2,595,4841996 238,236 7% 343,748 9% 289,587 8% 2,756,253 76% 3,627,8241997 265,269 8% 331,873 9% 261,446 7% 2,649,690 76% 3,508,2781998 999,749 25% 770,925 19% 383,323 9% 1,926,149 47% 4,080,1461999 862,337 17% 861,181 17% 301,980 6% 3,076,727 60% 5,102,2252000 519,006 14% 378,486 10% 283,090 8% 2,426,960 67% 3,607,5422001 808,369 18% 565,293 13% 245,176 5% 2,874,422 64% 4,493,260

45

Table B3. Virginia landings of Atlantic croaker from haul seine, by market grade and year (1973-2003). Included are percentages of total haul seine landings, by market grade.

YEAR Jumb

o % Large % Medium % Small % Unclassifi

ed % Total

Pounds

Total Pounds Total Pounds Total

Pounds Total

Pounds Total Pounds

1973 442,201 100% 442,2011974 99,716 100% 99,7161975 394,682 100% 394,6821976 625,722 100% 625,7221977 930,088 100% 930,0881978 665,752 100% 665,7521979 403,315 100% 403,3151980 45,655 100% 45,6551981 64,158 100% 64,1581982 188 100% 1881983 10,596 100% 10,5961984 177,621 100% 177,6211985 1,110,437 100% 1,110,4371986 1,302,181 100% 1,302,1811987 1,778,874 100% 1,778,8741988 694,972 100% 694,9721989 2,475 1% 101,327 36% 122,836 44% 52,182 19% 278,8201990 2,475 15% 98 1% 6,124 38% 7,352 46% 16,0491991 2,475 13.5% 153 1% 15,654 85% 83 0% 18,3651992 4% 34,068 8% 403,818 89% 13,624 3% 451,5101993 17,830 1% 169,145 20% 472,564 56% 186,056 22% 845,5951994 8,879 2% 196,214 18% 477,394 45% 384,130 36% 1,066,6171995 18,965 11% 362,295 28% 422,373 33% 476,274 37% 1,279,9071996 136,670 16% 599,442 39% 195,490 13% 599,818 39% 1,531,4201997 244,199 17% 1,059,130 33% 528,106 17% 1,338,487 42% 3,169,9221998 552,139 36% 838,135 35% 280,575 12% 726,940 30% 2,397,7891999 871,837 8% 784,507 28% 253,320 9% 928,186 33% 2,837,8502000 222,963 9% 486,269 23% 531,970 25% 890,721 42% 2,131,9232001 197,843 2% 810,614 37% 376,472 17% 833,221 38% 2,218,1502002 8,460 0.3% 49,109 2% 1,028,992 38% 279,576 10% 1,360,777 50% 2,726,9142003 847 0.03% 66,999 2% 1,598,710 53% 174,012 6% 1,200,188 39% 3,040,756

46

Table B4. Estimated Scrap Landings from Virginia in metric tons

Year Scrap estimates 1973 119.36 1974 101.22 1975 341.81 1976 464.71 1977 767.32 1978 736.51 1979 202.79 1980 64.29 1981 44.51 1982 8.64 1983 13.16 1984 77.38 1985 248.85 1986 285.27 1987 352.29 1988 183.66 1989 44.64 1990 14.06 1991 14.69 1992 141.62 1993 330.56 1994 266.51 1995 193.88 1996 48.40 1997 103.20 1998 66.40 1999 31.10 2000 19.80 2001 24.14 2002 15.82

47

Table B5. Size distribution of Virginia scrap component in 20 mm intervals Size class (mm) Year 120 140 160 180 200 220 240 Grand total 1989 0 0 3,815 7,696 90,966 323,070 57,921 483,468 1990 0 0 100 3,058 42,342 85,127 13,592 144,219 1991 64 64 0 8,385 44,054 95,715 14,922 163,204 1992 0 0 854 23,375 347,810 1,104,836 286,543 1,763,418 1993 0 0 0 12,606 430,501 2,664,577 1,009,569 4,117,253 1994 0 0 24,407 230,550 270,564 2,229,771 754,107 3,509,399 1995 0 0 29,422 366,644 705,240 1,156,509 265,687 2,523,502 1996 0 0 0 2,498 22,371 294,867 197,647 517,383 1997 0 0 0 0 108,811 799,504 321,222 1,229,537 1998 0 0 0 2,030 30,485 499,216 244,078 775,809 1999 0 0 0 0 26,377 207,362 149,219 382,958 2000 0 3,861 0 0 10,886 130,101 102,988 247,836 2001 0 0 0 0 2,493 91,558 99,377 193,428 2002 0 0 0 41,042 24,542 86,583 41,674 193,841 Grand total 64 3,925 58,598 697,884 2,157,442 9,768,796 3,558,546 16,245,255

48

APPENDIX C: Estimates of annual Atlantic croaker bycatch in the North Carolina shrimp trawl fishery, 1973-2002, based on a simple fish: shrimp ratio approach Annual estimates of Atlantic croaker bycatch-at-age in the North Carolina shrimp trawl fishery were produced for 1973 through 2002. Given the lack of detailed effort data and limited bycatch characterization data, estimates were produced using a fish catch to shrimp catch ratio method. Annual estimates in weight were converted to bycatch-at-age in numbers using length frequencies from the available bycatch data and the length-weight relationship and growth model derived from the North Carolina Division of Marine Fisheries (NCDMF) dataset in the current assessment. Ages present in the bycatch estimates ranged from 0 to 2. Over 99% of bycatch was age 0, which equated to 264 million fish, in 1994 (most reliable estimate in time series). Annual bycatch was on the order of millions of fish for age 1 and tens of thousands for age 2. These estimates must be considered extremely crude. Data sources: NC commercial shrimp trawl landings, 1973-2002 Annual landings, in pounds, from the shrimp trawl fishery were provided by NCDMF for Atlantic croaker and penaeid shrimp from 1973-2002. Over the time series, shrimp landings have averaged 6.5 million pounds ranging from 2.4 million (1981) to 11.4 million (1985) with no strong temporal trend (Figure C1). Catches made in Inside waters (sounds, coastal rivers, etc.) account for approximately 75% of annual shrimp landings, with the remaining 25% coming from Ocean waters. From 1973-2002, landings of Atlantic croaker from the NC shrimp trawl fishery have averaged 208,000 pounds. Landings declined steadily from 820,336 pounds in 1982, to 1,693 pounds in 2002 (Figure C2). In marked contrast to the overall decline, landings in 1996 and 1997 were both over 500,000 pounds. The spike in landings was attributed to a brief change in fishing behavior that was quickly modified by regulation (Tina Moore, pers. comm.1). Wolff, 1972 (Wolff) From June through August 1970, 39 trawl tows were sampled to determine discard ratios of finfish to commercially valuable shrimp by weight. Of the 39 tows, 4 were classified as “Ocean”, 18 as “Core Sound”, and 17 as “Pamlico Sound”. In addition to general location, day vs. night, total finfish catch weight, total shrimp catch weight, and the resulting fish:shrimp ratio were reported for each sampled trawl tow. Finfish species composition and percent by weight were reported for all tows combined. No length data were available from the 39 tows.

1 Tina Moore, North Carolina Division of Marine Fisheries 943 Washington Square Mall, Washington, NC 27889

49

NMFS bycatch characterization, 1992-1994 (NMFS) From 1992-1994, approximately 685 trawl tows were sampled during a NMFS bycatch characterization study. Data available from each sample included location, tow duration, gear information, total weight of penaeid shrimp by species, total weight and total number of Atlantic croaker. Lengths (TL mm) were recorded for Atlantic croaker from approximately 288 tows. Of the 685 tows, 17 were made in 1992, 146 in 1993, and 522 in 1994. By area, 36 were in Ocean waters, 629 were in Inside waters, and 20 had missing or erroneous location information. These data are summarized in Nance et al., 1997. Johnson, 2003 (Johnson) From June, 1999 through July, 2000, 56 trawl tows were sampled during a University of North Carolina shrimp trawl discard study (Johnson, 2003). Data available from each sample included location, tow duration, gear information, total weight of penaeid shrimp by species, total weight and total number of Atlantic croaker. Lengths (TL mm) were recorded from the 54 tows that caught Atlantic croaker. By year, 34 tows were sampled in 1999, 22 in 2000. Tows were made in two Inside water areas: the Neuse River and Core Sound. No tows from Ocean waters were sampled during this study. Methods and estimates: Atlantic croaker:shrimp ratios For each of the three bycatch and discard datasets described above, Atlantic croaker:shrimp ratios were calculated by dividing Atlantic croaker catch weight summed across all tows by shrimp catch weight summed across all tows (Table C1). Since tow duration was not available from Wolff, differences in duration among tows were not taken into account. While not desirable, this decision was made to keep ratio estimation consistent among the three datasets. For the Wolff and NMFS datasets, ratios were also calculated by area: Inside and Ocean. For NMFS and Johnson, tow catches were summed across years. For Wolff, the fish:shrimp ratio for all 39 tows was 5.38:1. Atlantic croaker made up 24.2% by weight of the total finfish catch from all tows. 24.2% of 5.38 is approximately 1.30, so the overall Atlantic croaker:shrimp ratio was 1.30:1. The fish:shrimp ratio was 14.0:1 for Ocean tows pooled, 1.6:1 for Core Sound, and 12.5:1 for Pamlico Sound. For Core and Pamlico Sounds combined, the ratio was 3.9:1. As species composition and percent by weight were not available for each tow, Atlantic croaker:shrimp ratios could only be calculated by area using 24.2% reported for all tows combined. The resulting Atlantic croaker:shrimp ratios by area were 3.4:1 for Ocean, 0.4:1 for Core Sound, 3.0:1 for Pamlico Sound, and 0.9:1 for Inside waters (Core and Pamlico Sounds combined). For NMFS, the Atlantic croaker:shrimp ratio for all years, areas combined was 1.66:1. By area, the ratio was 0.25:1 for Ocean waters and 1.81:1 for Inside waters. While not

50

used in subsequent calculations, ratios by year (pooled over area) were 1.83:1 in 1992, 1.07:1 in 1993, and 1.77:1 in 1994. For Johnson, the Atlantic croaker:shrimp ratio for both years combined was 0.60:1. Ratios could not be calculated by area as all sampled tows were made in Inside waters. While not used in subsequent calculations, ratios by year were 0.37:1 in 1999, and 1.01:1 in 2000. The three ratios (one from each bycatch dataset) based on Atlantic croaker and shrimp catches pooled over years and areas were considered to be the base case for subsequent calculations. Ratios by area were calculated and carried forward as one possible alternative. Annual Atlantic croaker bycatch by weight The first step in estimating total annual bycatch of Atlantic croaker was deciding how to apply the ratios from the three datasets to the time series, 1973-2002. The Wolff ratio was used for years 1972 through 1991. The NMFS ratio was used for 1992-1998, and the Johnson ratio was used for 1999-2002. In this method, a ratio was used from the first year in which the underlying data were collected until the year preceding the next available ratio. There are serious shortcomings to this method, and numerous alternatives could be employed. This issue is revisited in the Discussion section. After allocating the years in the time series among the three ratios, annual Atlantic croaker bycatch was calculated by multiplying annual shrimp landings by the appropriate Atlantic croaker:shrimp ratio and then subtracting the reported Atlantic croaker landings:

yearyearyear ngsortedLandiratioShrimpCroaingsShrimpLandBycatchCroa Rep:kerker

with all landings from the NC commercial shrimp trawl fishery (Table C2). Annual Atlantic croaker bycatch from 1973-2002 averaged 8.04 million pounds with a range of 2.73 million in 1981 to 14.56 million in 1985. While there was no clear trend over the entire time series, there appeared to be a decline in bycatch estimates from the early 1990’s through 2002 (Figure C3a). Annual Atlantic croaker bycatch was estimated by area, as well. Bycatch estimates for Inside and Ocean waters were calculated separately and then summed to produce total annual estimates (Table C3). The NMFS Ocean waters Atlantic croaker:shrimp ratio was used for 1992-2002 since all Johnson samples were from Inside waters. The by area estimate generally exceeded the area pooled estimate from 1972-1991 (Figure C3a). The by area estimate was less than the area pooled estimate from 1992-2002. This switch is most readily explained by the large decrease in the Ocean waters Atlantic croaker:shrimp ratio from 3.40:1, 1972-1991, to 0.25:1 for 1992-2002.

51

Length Frequency Distributions Weighted length frequency distributions were calculated for the NMFS and Johnson datasets separately. Observed numbers-at-length from each sampled net were expanded to the tow level and then summed across all tows with Atlantic croaker catch to produce weighted length frequency distributions. Length frequency distributions were somewhat different between the two datasets (Figure C4). Both had a minimum length of about 30mm. However, the NMFS distribution had a mode of 120mm and a maximum of about 250mm compared to Johnson with a mode of 90mm and a maximum of about 180mm. For NMFS, separate length frequencies were calculated for Inside and Ocean waters. The Inside distribution ranged from 30mm to 250mm with a mode of 120mm (Figure C5). The Ocean distribution was truncated in comparison, with a range of 90-180mm and a mode of 110mm. Mean weight per fish Mean weight per fish was calculated for the NMFS and Johnson datasets separately. Lengths (mm) from the weighted length frequencies were converted to weights (kg) using the length weight relationship from the current assessment: Weight = a(Length)b, where a = 5.49 x 10-9, b = 3.13. Individual weights were then summed and divided by the total number to produce a mean. The mean weight for NMFS was 0.02 kg (0.043 lbs.). Mean weight for Johnson was smaller at 0.013 kg (0.028 lbs.). Mean weights were also calculated by area for NMFS. For Ocean waters, the mean was 0.021 kg (0.047 lbs.). Inside waters mean was slightly smaller at 0.02 kg (0.044 lbs.). Annual Atlantic Croaker Bycatch By Number To convert bycatch from weight to numbers, annual Atlantic croaker bycatch by weight was divided by mean weight per fish. The NMFS mean weight estimate was used for 1973-1998 since there was no length or individual fish weight information available from Wolff. Johnson mean weight was used for 1999-2002. As landings were in pounds, mean weight estimates were converted from kilograms to pounds prior to calculations. Estimates of bycatch by number were two orders of magnitude greater than estimates by weight. The average number of Atlantic croaker in shrimp trawl bycatch over the time series was 194 million, annually, ranging from 63.2 million, 1981, to 337 million, 1985 (Table C2). There was a similar decline in bycatch estimates by number from the early 1990’s through 2002 (Figure C3b). However, the decline in numbers did not exactly

52

match the decline in weight due to the smaller mean weight per fish estimate from Johnson used in years 1999-2002. Annual bycatch in numbers was also estimated by area. As before, estimates were produced separately for Ocean and Inside waters, then summed to produce total annual estimates. Mean weight per fish for Ocean waters from NMFS was used for 1973 through 2002. For Inside waters, NMFS was used for 1973-1998, and Johnson was used for 1999-2002. Calculated by area, average annual bycatch in numbers was 192 million, 1973-2002 (Table C3). As with weight estimates, bycatch in numbers by area were generally greater than the area pooled estimates from 1973-1991 and less than area pooled from 1992-2002 (Figure C3b). Age Composition For NMFS and Johnson, age compositions were produced by converting the weighted numbers-at-length, pooled over years and areas, to numbers-at-age using the growth model from the current assessment. The von Bertalanffy equation: L = L(1-exp(-k(Age – to))), where L = 434.6mm, k = 0.2415, to = -1.9572, was rearranged to solve for age: Age = ( ln( 1 – ( L / L ) ) ) / (-k) + to. Calculated ages less than zero were set to age 0. Non-negative ages were rounded to the nearest whole age. Age compositions for both NMFS and Johnson were comprised primarily or entirely of age 0 fish. For NMFS, age 0 made up the overwhelming majority at just over 99.32%. Age 1 made up about 0.67%, and age 2 made up slightly more than 0.01%. Age 0 fish made up 100% of the Johnson age composition. Age compositions by area could only be calculated for NMFS. Inside waters age composition was virtually identical to the area pooled composition: 99.32% - age 0, 0.67% - age 1, 0.01% - age 2. Ocean waters fish were 100% age 0. Bycatch-at-age Annual estimates of bycatch by number were multiplied by each age percentage in the appropriate age composition to produce bycatch-at-age in numbers. NMFS age composition was used for years 1973-1998. Johnson was used for 1999-2002. The resulting estimates are given in Table C4 and Figure C6. The average number of age 0 fish caught annually was approximately 193 million ranging from 62.8 million, in 1981, to 334.8 million, 1985. Age 1 and age 2 annual bycatch in numbers averaged 1.14 million and 19,130, respectively. However, it must be noted that ages 1 and 2 were

53

absent from estimates for years 1999-2002 due to the smaller length frequency distribution of Johnson. As before, this issue will be expanded upon in the Discussion section. While overall patterns in bycatch-at-age calculated by area were very similar to area pooled, annual estimates by area were consistently smaller for ages 1 and 2 (Table C5, Figure C6). The annual averages for ages 1 and 2 were 734,080 and 12,273, respectively, both considerably less than the area pooled averages of 1.14 million and 19,130. Conversely, age 0 averaged 191 million annually which was quite similar to the area pooled average of 193 million. As with area pooled, ages 1 and 2 were absent from the by area estimates for years 1999-2002. For these years, the Johnson age composition (100% age 0) was used for Inside and the NMFS Ocean waters age composition (100% age 0) was used for Ocean. Discussion: Due to the scarcity of information concerning Atlantic croaker bycatch in the North Carolina shrimp trawl fishery relative to the time series of the current assessment, numerous subjective decisions were made to produce this initial set of estimates. The rationale for, along with possible alternatives to, these decisions are provided below. Undoubtedly, significant changes will need to be made to the methodology and resulting estimates presented in this report. Fish: Shrimp ratio bycatch estimation approach At the heart of this approach are at least two key assumptions. First, Atlantic croaker abundance and shrimp abundance are related, or more correctly, the catchability of Atlantic croaker in the North Carolina shrimp trawl fishery and the catchability of shrimp in said fishery are directly, linearly related. The second assumption is that available bycatch information is sufficient to produce ratio estimates representative of the fishery over the time series considered. It is beyond the capabilities of the author to address these assumptions other than to provide several references on the subject (Peuser 1996; Nance et al. 1997; Diamond 2003) and to state that 6 years of bycatch characterization data (5 of which are from 1992-2000) are being applied to a 30 year time series. Ratio calculations One of the goals in producing these initial estimates was to incorporate all bycatch information that was readily available. Because the three bycatch datasets had different levels of detail, all methods and estimates were standardized to the lowest level. The Wolff dataset had the lowest level of detail providing most information at the tow level (general area, total weight of shrimp and total weight of fish per trawl tow) and one critical piece of information at the study level (proportional fish species composition of total fish landings summed over all tows). Shrimp and Atlantic croaker catches from NMFS and Johnson datasets were expanded, as needed, to the tow level and then summed across all tows to produce a base case Atlantic croaker:shrimp ratio for each of

54

the two datasets consistent with the Wolff base case ratio. This method ignores all ancillary information from NMFS and Johnson that could have been used to calculate ratios by strata such as year or season, based on catches standardized to a consistent unit of effort (e.g. tow hour). The “by area” method was one alternative to pooling all information for each study. Unfortunately, this method had serious flaws. Only Wolff and NMFS datasets had observations from Ocean waters, and of those Wolff had only four. Since there were no Ocean tows in the Johnson dataset, the NMFS Ocean ratio was used for 1999-2002. Producing Atlantic croaker:shrimp ratios by area for Wolff required the assumption that the Atlantic croaker proportion of total fish landings (0.242) was not significantly different from what that proportion might have been for Inside and Ocean tows considered separately. Discards vs. Landings Reported landings of Atlantic croaker must be considered when producing bycatch estimates. Sampling in the three bycatch studies was conducted at sea, meaning that any ratios calculated from these data would reflect Atlantic croaker to be discarded as well as Atlantic croaker to be landed. For this reason, annual reported landings from the shrimp trawl fishery were subtracted from the total bycatch estimate to produce a discard bycatch estimate. This method assumes that reported landings come from the total bycatch indiscriminately. It is more likely that reported landings are comprised of the largest fish in the bycatch, disproportionate to their numbers. If this were the case, the bycatch-at-age estimates presented here would be biased high for Ages 1 and 2, and low for Age 0. Some other method for allocating the reported landings by size or age group should be evaluated. Length Information Length information was only available from NMFS and Johnson datasets. For this reason, mean size and age composition from NMFS were used in calculations for years 1973-1991 when the Wolff ratio was used to produce the annual bycatch estimate by weight. Starting in 1992, the first year of the NMFS dataset, BRD’s were being implemented in North Carolina. Undoubtedly, BRD’s have affected the size distribution of Atlantic croaker present in bycatch. However, this size distribution has to be applied to years prior to BRD implementation. Possible Alternatives The following paragraphs provide alternative ratio approaches using the current datasets, with advantages, disadvantages and potential changes in the estimates relative to the base case, area pooled.

55

Consider only NMFS dataset: This is the most extensive bycatch characterization dataset currently available. It includes hundreds of observed tows providing the largest spatial and temporal coverage. The NMFS mean size and age composition are already being applied to most of the time series, 1973-1998. Disadvantages include applying three consecutive years of data to the remaining 27 and applying a ratio based on BRD impacted catches to years prior to BRD implementation. Likely changes to the estimates include ages 1 and 2 being present for years 1999-2002 and a slight increase in annual bycatch overall. Pool all datasets: Given the limited information available and realizing that over a 30 year time series many aspects of the Atlantic croaker population(s), shrimp population(s), and the shrimp fishery are subject to change, pooling all available information might produce an average set of estimates for the time series. This approach would require some weighting scheme among the datasets or the resulting estimates would still be dominated by the NMFS. Pooling all datasets will not address the lack of length distribution data prior to 1992 and BRD implementation. Effects on bycatch estimates would depend heavily on the weighting scheme with the exception that Ages 1 and 2 are likely to be present for years 1999-2002. Smooth transitions between datasets: The current stepwise approach produces dramatic changes across the time series, most notably the sudden disappearance of ages 1 and 2 from bycatch in 1999, continuing through 2002. A smoothing function would allow for less abrupt changes that might be more realistic. This approach would not address the lack of length information prior to 1992 and would require some means of evaluating the smoothing function. Ages 1 and 2 would likely reappear for some portion of the 1999-2002 time series with other changes being less noticeable. Calculate ratios using different methodologies appropriate to the level of coverage in each dataset: This approach might improve estimates for the latter part of the time series, 1992-2002, as separate ratios could be calculated for more spatial, temporal strata. Coverage in terms of length information would have to be evaluated, and two-thirds of the time series will still have relatively less precise estimates based on length data impacted by BRD’s. Likely impacts on the estimates would be minimal prior to 1992. Literature Cited: Diamond, S.L. 2003. Estimation of bycatch in shrimp trawl fisheries: a comparison of estimation methods using field data and simulated data. Fish. Bull. 101:484-500. Johnson, G. 2003. The role of trawl discards in sustaining blue crab production. Fishery Resource Grant 99-EP-07. 127 pp. Nance, J., E. Scott-Denton, E. Martinez, J. Watson, A. Shah, and D. Foster. 1997. Bycatch in the southeast shrimp trawl fishery. A data summary report, SFA Task N-10.03.

56

Peuser, R. (ed.) 1996. Estimates of finfish bycatch in the South Atlantic shrimp fishery. Prepared by the SEAMAP South Atlantic Committee, Shrimp Bycatch Work Group, Final Report. Atlantic States Marine Fisheries Commission, April 1996, 64 pp. Wolff, M. 1972. A study of North Carolina scrap fishery. NC Department of Natural and Economic Resources, Special Scientific Report 20, 29 pp.

57

Table C1. Summary of Atlantic croaker:shrimp ratios by weight. Atlantic croaker and shrimp total weights are in pounds for Wolff, and kilograms for NMFS and Johnson. Sample sizes for NMFS Inside and NMFS Ocean do not sum to NMFS Pooled as 20 observations had missing tow coordinates.

Dataset WOLFF WOLFF WOLFF NMFS NMFS NMFS JOHNSONN 39 35 4 685 629 36 56Area POOLED INSIDE OUTSIDE POOLED INSIDE OUTSIDE INSIDE ONLYAtlantic croaker total weight 265.97 164.25 101.72 23832.69 23367.48 264.22 586.38Shrimp Total Weight 204.33 174.33 30.00 14294.80 12915.34 1061.86 976.24C:S Ratio 1.30 0.94 3.39 1.67 1.81 0.25 0.60

58

Table C2. Reported annual shrimp and Atlantic croaker landings from the North Carolina shrimp trawl fishery, with estimated Atlantic croaker bycatch and number. C:S ratios and Atlantic croaker mean weights were calculated using the area pooled approach.

DiscardAtlantic croaker Discard

ShrimpAtlantic croaker Total Atlantic croaker mean Atlantic croaker landed landed croaker bycatch Weight bycatchYear (lbs.) (lbs.) C:S ratio (lbs.) (lbs.) (lbs) (No.)1970 4,918,000 34,500 1.30 6,401,552 6,367,052 0 147,412,0691971 7,408,900 29,700 1.30 9,643,850 9,614,150 0 222,589,9631972 5,445,100 73,001 1.30 7,087,655 7,014,654 0 162,405,5781973 4,864,700 163,146 1.30 6,332,173 6,169,027 0 142,827,3441974 8,227,700 255,596 1.30 10,709,648 10,454,052 0 242,035,6311975 4,962,200 330,061 1.30 6,459,085 6,129,024 0 141,901,1721976 6,490,500 137,927 1.30 8,448,408 8,310,481 0 192,406,9851977 5,578,900 254,361 1.30 7,261,817 7,007,456 0 162,238,9231978 2,880,331 188,699 1.30 3,749,204 3,560,505 0 82,433,9911979 4,613,389 376,698 1.30 6,005,052 5,628,354 0 130,309,5091980 9,210,911 447,142 1.30 11,989,452 11,542,310 0 267,231,3361981 2,432,165 435,374 1.30 3,165,846 2,730,472 0 63,216,7801982 6,666,224 820,336 1.30 8,677,141 7,856,805 0 181,903,3171983 5,884,672 425,522 1.30 7,659,828 7,234,306 0 167,491,0151984 4,682,496 105,667 1.30 6,095,006 5,989,339 0 138,667,1441985 11,397,253 277,047 1.30 14,835,320 14,558,273 0 337,057,9131986 5,969,024 215,283 1.30 7,769,625 7,554,342 0 174,900,6031987 4,207,170 80,473 1.30 5,476,294 5,395,821 0 124,925,8281988 7,869,873 118,500 1.30 10,243,879 10,125,379 0 234,426,1041989 8,643,154 205,161 1.30 11,250,426 11,045,265 0 255,723,6071990 7,538,761 101,216 1.30 9,812,885 9,711,669 0 224,847,7441991 10,163,807 114,765 1.30 13,229,796 13,115,031 0 303,643,4951992 5,200,780 36,319 1.67 8,670,889 8,634,570 0 199,910,4001993 6,144,215 40,836 1.67 10,243,811 10,202,975 0 236,222,6221994 6,893,428 14,821 1.67 11,492,920 11,478,100 0 265,744,7251995 7,911,321 19,013 1.67 13,189,981 13,170,968 0 304,938,5731996 4,876,299 505,599 1.67 8,129,905 7,624,306 0 176,520,4361997 6,451,887 549,275 1.67 10,756,770 10,207,495 0 236,327,2841998 4,271,323 9,197 1.67 7,121,272 7,112,075 0 164,661,0891999 8,109,944 6,987 0.60 4,871,247 4,864,260 0 174,299,2492000 9,443,835 1,180 0.60 5,672,450 5,671,270 0 203,216,5662001 4,747,112 2,257 0.60 2,851,358 2,849,101 0 102,090,8002002 8,834,301 1,693 0.60 5,306,333 5,304,640 0 190,079,219

59

Table C3. Reported annual shrimp and Atlantic croaker landings from the North Carolina shrimp trawl fishery, with estimated Atlantic croaker bycatch by weight and number. C:S ratios and Atlantic croaker mean weights were calculated using the alternative by area approach.

INSIDE Discarded Atlantic croaker Discarded Shrimp Atlantic croaker Total Atlantic croaker mean Atlantic croaker landed landed Atlantic croaker bycatch weight bycatchYear (lbs.) (lbs.) C:S Ratio (lbs.) (lbs.) (lbs) (No.)

1972 3,326,912 25,483 1 3,134,458 3,108,975 0.0435 71,462,820 1973 2,975,106 83,241 1 2,803,003 2,719,762 0.0435 62,516,383 1974 6,258,665 146,232 1 5,896,615 5,750,383 0.0435 132,178,181 1975 3,051,935 116,365 1 2,875,387 2,759,022 0.0435 63,418,827 1976 4,635,186 75,168 1 4,367,051 4,291,883 0.0435 98,653,124 1977 4,667,283 145,440 1 4,397,291 4,251,851 0.0435 97,732,955 1978 1,787,660 169,967 1 1,684,248 1,514,281 0.0435 34,807,226 1979 3,121,182 297,881 1 2,940,629 2,642,748 0.0435 60,746,138 1980 7,646,578 381,136 1 7,204,241 6,823,105 0.0435 156,835,729 1981 1,929,037 407,151 1 1,817,447 1,410,296 0.0435 32,417,022 1982 5,087,026 772,272 1 4,792,753 4,020,481 0.0435 92,414,685 1983 4,283,091 374,028 1 4,035,324 3,661,296 0.0435 84,158,463 1984 3,025,281 80,307 1 2,850,275 2,769,968 0.0435 63,670,429 1985 10,016,284 211,253 1 9,436,864 9,225,611 0.0435 212,059,697 1986 4,768,243 160,096 1 4,492,411 4,332,315 0.0435 99,582,492 1987 3,035,150 46,482 1 2,859,573 2,813,091 0.0435 64,661,655 1988 6,015,313 85,970 1 5,667,341 5,581,371 0.0435 128,293,260 1989 6,569,740 171,442 1 6,189,695 6,018,253 0.0435 138,335,435 1990 6,155,089 81,536 1 5,799,031 5,717,495 0.0435 131,422,210 1991 8,812,513 93,316 1 8,302,729 8,209,413 0.0435 188,701,380 1992 4,232,962 23,020 2 7,658,616 7,635,596 0.0435 175,511,639 1993 4,344,978 7,471 2 7,861,284 7,853,813 0.0435 180,527,578 1994 5,242,490 8,392 2 9,485,136 9,476,744 0.0435 217,832,226 1995 5,729,196 11,735 2 10,365,724 10,353,989 0.0435 237,996,566 1996 3,054,886 6,620 2 5,527,146 5,520,527 0.0435 126,894,704 1997 4,911,723 4,152 2 8,886,685 8,882,533 0.0435 204,173,698 1998 2,019,600 8,031 2 3,654,023 3,645,992 0.0435 83,806,694 1999 5,275,158 1,744 1 3,168,530 3,166,786 0.0279 113,474,275 2000 7,847,702 999 1 4,713,731 4,712,732 0.0279 168,869,623 2001 3,493,218 2,161 1 2,098,205 2,096,044 0.0279 75,106,797 2002 7,511,154 1,083 1 4,511,583 4,510,500 0.0279 161,623,104

60

Table C3. Continued.

OCEAN Total Total Discarded Atlantic croaker Discarded discard discard Shrimp Atlantic croaker Total Atlantic croaker mean Atlantic croaker Atlantic croaker Atlantic croaker landed landed Atlantic croaker bycatch weight bycatch bycatch bycatch

Year (lbs.) (lbs.) C:S ratio (lbs.) (lbs.) (lbs) (No.) (lbs.) (No.)1972 2,118,190 47,518 3 7,182,237 7,134,719 0.0473 150,935,553 10,243,694 222,398,374 1973 1,889,582 79,905 3 6,407,086 6,327,181 0.0473 133,852,025 9,046,943 196,368,408 1974 1,969,214 109,364 3 6,677,098 6,567,734 0.0473 138,940,935 12,318,117 271,119,115 1975 1,910,234 213,696 3 6,477,112 6,263,416 0.0473 132,503,063 9,022,438 195,921,890 1976 1,855,285 62,759 3 6,290,794 6,228,035 0.0473 131,754,577 10,519,918 230,407,701 1977 911,933 108,921 3 3,092,130 2,983,209 0.0473 63,110,027 7,235,060 160,842,982 1978 1,093,067 18,732 3 3,706,309 3,687,577 0.0473 78,010,983 5,201,858 112,818,210 1979 1,494,932 78,817 3 5,068,930 4,990,113 0.0473 105,566,232 7,632,860 166,312,370 1980 1,580,401 66,006 3 5,358,733 5,292,727 0.0473 111,968,064 12,115,832 268,803,793 1981 505,898 28,223 3 1,715,370 1,687,147 0.0473 35,691,728 3,097,442 68,108,749 1982 1,601,111 48,064 3 5,428,955 5,380,891 0.0473 113,833,186 9,401,372 206,247,871 1983 1,601,956 51,494 3 5,431,821 5,380,327 0.0473 113,821,237 9,041,622 197,979,700 1984 1,647,793 25,360 3 5,587,242 5,561,882 0.0473 117,662,058 8,331,850 181,332,486 1985 1,381,502 65,794 3 4,684,318 4,618,524 0.0473 97,705,239 13,844,135 309,764,936 1986 1,173,063 55,187 3 3,977,555 3,922,368 0.0473 82,978,003 8,254,683 182,560,494 1987 1,168,458 33,991 3 3,961,940 3,927,949 0.0473 83,096,082 6,741,041 147,757,738 1988 1,826,134 32,530 3 6,191,950 6,159,420 0.0473 130,303,031 11,740,791 258,596,291 1989 2,073,275 33,719 3 7,029,942 6,996,223 0.0473 148,005,655 13,014,476 286,341,090 1990 1,381,184 19,680 3 4,683,240 4,663,560 0.0473 98,657,973 10,381,054 230,080,183 1991 1,341,367 21,449 3 4,548,230 4,526,781 0.0473 95,764,420 12,736,194 284,465,800 1992 967,602 13,299 0 240,766 227,467 0.0473 4,812,087 7,863,063 180,323,726 1993 1,799,000 33,365 0 447,641 414,276 0.0473 8,764,045 8,268,090 189,291,623 1994 1,643,201 6,429 0 408,874 402,445 0.0473 8,513,760 9,879,189 226,345,986 1995 2,181,634 7,278 0 542,851 535,573 0.0473 11,330,091 10,889,562 249,326,657 1996 1,811,952 498,980 0 450,864 - 0.0473 - 5,520,527 126,894,704 1997 1,539,725 545,123 0 383,126 - 0.0473 - 8,882,533 204,173,698 1998 2,250,604 1,166 0 560,013 558,847 0.0473 11,822,446 4,204,839 95,629,141 1999 2,834,540 5,243 0 705,312 700,069 1.0473 668,471 3,866,855 114,142,745 2000 1,596,695 181 0 397,302 397,121 2.0473 193,976 5,109,853 169,063,599 2001 1,256,792 96 0 312,725 312,629 3.0473 102,593 2,408,673 75,209,390 2002 1,370,572 610 0 341,036 340,426 4.0473 84,113 4,850,926 161,707,217

61

Table C4. Estimates of annual croaker bycatch by age group, in weight and numbers,

from the North Carolina shrimp trawl fishery, 1973-2002. Estimates were calculated using the area pooled approach.

Bycatch at age (Number) Discard Atlantic croaker bycatch Age proportions Bycatch at age Year (No.) Age0 Age1 Age2 Age0 Age1 Age21972 162,405,578 0.993226 0.006662 0.000111 161,305,482 1,082,007 18,0901973 142,827,344 0.993226 0.006662 0.000111 141,859,866 951,569 15,9091974 242,035,631 0.993226 0.006662 0.000111 240,396,139 1,612,532 26,9601975 141,901,172 0.993226 0.006662 0.000111 140,939,968 945,399 15,8061976 192,406,985 0.993226 0.006662 0.000111 191,103,666 1,281,887 21,4321977 162,238,923 0.993226 0.006662 0.000111 161,139,955 1,080,896 18,0711978 82,433,991 0.993226 0.006662 0.000111 81,875,603 549,206 9,1821979 130,309,509 0.993226 0.006662 0.000111 129,426,823 868,171 14,5151980 267,231,336 0.993226 0.006662 0.000111 265,421,175 1,780,395 29,7661981 63,216,780 0.993226 0.006662 0.000111 62,788,565 421,174 7,0421982 181,903,317 0.993226 0.006662 0.000111 180,671,147 1,211,908 20,2621983 167,491,015 0.993226 0.006662 0.000111 166,356,471 1,115,888 18,6561984 138,667,144 0.993226 0.006662 0.000111 137,727,846 923,852 15,4461985 337,057,913 0.993226 0.006662 0.000111 334,774,763 2,245,606 37,5441986 174,900,603 0.993226 0.006662 0.000111 173,715,868 1,165,253 19,4821987 124,925,828 0.993226 0.006662 0.000111 124,079,611 832,303 13,9151988 234,426,104 0.993226 0.006662 0.000111 232,838,158 1,561,834 26,1121989 255,723,607 0.993226 0.006662 0.000111 253,991,396 1,703,726 28,4841990 224,847,744 0.993226 0.006662 0.000111 223,324,680 1,498,020 25,0451991 303,643,495 0.993226 0.006662 0.000111 301,586,686 2,022,986 33,8221992 199,910,400 0.993226 0.006662 0.000111 198,556,255 1,331,878 22,2671993 236,222,622 0.993226 0.006662 0.000111 234,622,507 1,573,803 26,3121994 265,744,725 0.993226 0.006662 0.000111 263,944,634 1,770,491 29,6011995 304,938,573 0.993226 0.006662 0.000111 302,872,992 2,031,615 33,9661996 176,520,436 0.993226 0.006662 0.000111 175,324,729 1,176,045 19,6621997 236,327,284 0.993226 0.006662 0.000111 234,726,459 1,574,501 26,3241998 164,661,089 0.993226 0.006662 0.000111 163,545,715 1,097,034 18,3411999 174,299,249 1 0 0 174,299,249 0 02000 203,216,566 1 0 0 203,216,566 0 02001 102,090,800 1 0 0 102,090,800 0 02002 190,079,219 1 0 0 190,079,219 0 0

62

Table C4. Continued. Bycatch at age (lbs.) Discard Atlantic croaker bycatch Age proportions bycatch at age Year (lbs.) Age0 Age1 Age2 Age0 Age1 Age21972 7,014,654 0.993226 0.006662 0.000111 6,967,139 46,734 7811973 6,169,027 0.993226 0.006662 0.000111 6,127,240 41,100 6871974 10,454,052 0.993226 0.006662 0.000111 10,383,238 69,649 1,1641975 6,129,024 0.993226 0.006662 0.000111 6,087,507 40,834 6831976 8,310,481 0.993226 0.006662 0.000111 8,254,188 55,368 9261977 7,007,456 0.993226 0.006662 0.000111 6,959,989 46,686 7811978 3,560,505 0.993226 0.006662 0.000111 3,536,387 23,721 3971979 5,628,354 0.993226 0.006662 0.000111 5,590,229 37,498 6271980 11,542,310 0.993226 0.006662 0.000111 11,464,125 76,899 1,2861981 2,730,472 0.993226 0.006662 0.000111 2,711,976 18,191 3041982 7,856,805 0.993226 0.006662 0.000111 7,803,584 52,345 8751983 7,234,306 0.993226 0.006662 0.000111 7,185,302 48,198 8061984 5,989,339 0.993226 0.006662 0.000111 5,948,769 39,903 6671985 14,558,273 0.993226 0.006662 0.000111 14,459,659 96,993 1,6221986 7,554,342 0.993226 0.006662 0.000111 7,503,171 50,330 8411987 5,395,821 0.993226 0.006662 0.000111 5,359,271 35,949 6011988 10,125,379 0.993226 0.006662 0.000111 10,056,792 67,459 1,1281989 11,045,265 0.993226 0.006662 0.000111 10,970,447 73,588 1,2301990 9,711,669 0.993226 0.006662 0.000111 9,645,884 64,703 1,0821991 13,115,031 0.993226 0.006662 0.000111 13,026,193 87,377 1,4611992 8,634,570 0.993226 0.006662 0.000111 8,576,082 57,527 9621993 10,202,975 0.993226 0.006662 0.000111 10,133,862 67,976 1,1361994 11,478,100 0.993226 0.006662 0.000111 11,400,350 76,471 1,2791995 13,170,968 0.993226 0.006662 0.000111 13,081,751 87,750 1,4671996 7,624,306 0.993226 0.006662 0.000111 7,572,661 50,796 8491997 10,207,495 0.993226 0.006662 0.000111 10,138,352 68,006 1,1371998 7,112,075 0.993226 0.006662 0.000111 7,063,899 47,383 7921999 4,864,260 1.000000 0.000000 0.000000 4,864,260 0 02000 5,671,270 1.000000 0.000000 0.000000 5,671,270 0 02001 2,849,101 1.000000 0.000000 0.000000 2,849,101 0 02002 5,304,640 1.000000 0.000000 0.000000 5,304,640 0 0

63

Table C5. Estimates of annual Atlantic croaker bycatch by age group, by number and by weight, from the North Carolina shrimp trawl fishery. Estimates were calculated using the alternative by area approach.

Inside Ocean Total Age proportions Bycatch at age (No.) Age proportions Bycatch at age (No.) Bycatch at age (No.)

Year

Discards Atlantic croaker bycatch

(No.) Age0 Age1 Age2 Age0 Age1 Age2

DiscardsAtlanticcroakerbycatch

(No.) Age0 Age1 Age2 Age0 Age1 Age2 Age0 Age1 Age2

1972 71,462,820 0.9929 0.0069 0.0001 70,957,955 496,563 8,302 150,935,553 1 0 0 150,935,553 0 0 221,893,509 496,563 8,302

1973 62,516,383 0.9929 0.0069 0.0001 62,074,722 434,398 7,263 133,852,025 1 0 0 133,852,025 0 0 195,926,747 434,398 7,263

1974 132,178,181 0.9929 0.0069 0.0001 131,244,378 918,447 15,355 138,940,935 1 0 0 138,940,935 0 0 270,185,313 918,447 15,355

1975 63,418,827 0.9929 0.0069 0.0001 62,970,790 440,669 7,367 132,503,063 1 0 0 132,503,063 0 0 195,473,854 440,669 7,367

1976 98,653,124 0.9929 0.0069 0.0001 97,956,166 685,497 11,461 131,754,577 1 0 0 131,754,577 0 0 229,710,743 685,497 11,461

1977 97,732,955 0.9929 0.0069 0.0001 97,042,499 679,103 11,354 63,110,027 1 0 0 63,110,027 0 0 160,152,525 679,103 11,354

1978 34,807,226 0.9929 0.0069 0.0001 34,561,323 241,860 4,044 78,010,983 1 0 0 78,010,983 0 0 112,572,306 241,860 4,044

1979 60,746,138 0.9929 0.0069 0.0001 60,316,983 422,098 7,057 105,566,232 1 0 0 105,566,232 0 0 165,883,215 422,098 7,057

1980 156,835,729 0.9929 0.0069 0.0001 155,727,728 1,089,781 18,220 111,968,064 1 0 0 111,968,064 0 0 267,695,792 1,089,781 18,220

1981 32,417,022 0.9929 0.0069 0.0001 32,188,004 225,251 3,766 35,691,728 1 0 0 35,691,728 0 0 67,879,732 225,251 3,766

1982 92,414,685 0.9929 0.0069 0.0001 91,761,801 642,148 10,736 113,833,186 1 0 0 113,833,186 0 0 205,594,987 642,148 10,736

1983 84,158,463 0.9929 0.0069 0.0001 83,563,906 584,780 9,777 113,821,237 1 0 0 113,821,237 0 0 197,385,144 584,780 9,777

1984 63,670,429 0.9929 0.0069 0.0001 63,220,615 442,417 7,397 117,662,058 1 0 0 117,662,058 0 0 180,882,672 442,417 7,397

1985 212,059,697 0.9929 0.0069 0.0001 210,561,554 1,473,508 24,635 97,705,239 1 0 0 97,705,239 0 0 308,266,792 1,473,508 24,635

1986 99,582,492 0.9929 0.0069 0.0001 98,878,969 691,954 11,569 82,978,003 1 0 0 82,978,003 0 0 181,856,971 691,954 11,569

64

Table C5. Continued.

Inside Ocean Total

Age proportions Bycatch at age Age proportions Bycatch at age Bycatch at age

Year

Discards Atlantic croaker bycatch

(No.) Age0 Age1 Age2 Age0 Age1 Age2

DiscardsAtlanticcroakerbycatch

(No.) Age0 Age1 Age2 Age0 Age1 Age2 Age0 Age1 Age2

1987 64,661,655 0.9929 0.0069 0.0001 64,204,838 449,305 7,512 83,096,082 1 0 0 83,096,082 0 0 147,300,921 449,305 7,512

1988 128,293,260 0.9929 0.0069 0.0001 127,386,903 891,453 14,904 130,303,031 1 0 0 130,303,031 0 0 257,689,934 891,453 14,904

1989 138,335,435 0.9929 0.0069 0.0001 137,358,133 961,231 16,071 148,005,655 1 0 0 148,005,655 0 0 285,363,788 961,231 16,071

1990 131,422,210 0.9929 0.0069 0.0001 130,493,748 913,194 15,268 98,657,973 1 0 0 98,657,973 0 0 229,151,721 913,194 15,268

1991 188,701,380 0.9929 0.0069 0.0001 187,368,257 1,311,202 21,922 95,764,420 1 0 0 95,764,420 0 0 283,132,676 1,311,202 21,922

1992 175,511,639 0.9929 0.0069 0.0001 174,271,697 1,219,552 20,390 4,812,087 1 0 0 4,812,087 0 0 179,083,785 1,219,552 20,390

1993 180,527,578 0.9929 0.0069 0.0001 179,252,201 1,254,406 20,972 8,764,045 1 0 0 8,764,045 0 0 188,016,245 1,254,406 20,972

1994 217,832,226 0.9929 0.0069 0.0001 216,293,301 1,513,619 25,306 8,513,760 1 0 0 8,513,760 0 0 224,807,061 1,513,619 25,306

1995 237,996,566 0.9929 0.0069 0.0001 236,315,186 1,653,732 27,648 11,330,091 1 0 0 11,330,091 0 0 247,645,277 1,653,732 27,648

1996 126,894,704 0.9929 0.0069 0.0001 125,998,228 881,735 14,742 0 1 0 0 0 0 0 125,998,228 881,735 14,742

1997 204,173,698 0.9929 0.0069 0.0001 202,731,267 1,418,712 23,719 0 1 0 0 0 0 0 202,731,267 1,418,712 23,719

1998 83,806,694 0.9929 0.0069 0.0001 83,214,623 582,335 9,736 11,822,446 1 0 0 11,822,446 0 0 95,037,069 582,335 9,736

1999 113,474,275 1.0000 0.0000 0.0000 113,474,275 0 0 668,471 1 0 0 668,471 0 0 114,142,745 0 0

2000 168,869,623 1.0000 0.0000 0.0000 168,869,623 0 0 193,976 1 0 0 193,976 0 0 169,063,599 0 0

2001 75,106,797 1.0000 0.0000 0.0000 75,106,797 0 0 102,593 1 0 0 102,593 0 0 75,209,390 0 0

2002 161,623,104 1.0000 0.0000 0.0000 161,623,104 0 0 84,113 1 0 0 84,113 0 0 161,707,217 0 0

65

Table C5. Continued.

Inside Ocean Total

Age proportions Bycatch at age Age proportions Bycatch at age Bycatch at age

Year

Discarded Atlanticcroaker bycatch

(lbs.) Age0 Age1 Age2 Age0 Age1 Age2

DiscardedAtlantic croaker

bycatch(lbs.)Age0Age1Age2 Age0Age1Age2 Age0 Age1 Age2

1972 3,108,975 0.99290.0069 0.0001 3,087,011 21,603 361 7,134,719 1 0 0 7,134,719 0 0 10,221,730 21,603 361

1973 2,719,762 0.99290.0069 0.0001 2,700,547 18,898 316 6,327,181 1 0 0 6,327,181 0 0 9,027,729 18,898 316

1974 5,750,383 0.99290.0069 0.0001 5,709,758 39,957 668 6,567,734 1 0 0 6,567,734 0 0 12,277,492 39,957 668

1975 2,759,022 0.99290.0069 0.0001 2,739,531 19,171 321 6,263,416 1 0 0 6,263,416 0 0 9,002,947 19,171 321

1976 4,291,883 0.99290.0069 0.0001 4,261,562 29,822 499 6,228,035 1 0 0 6,228,035 0 0 10,489,597 29,822 499

1977 4,251,851 0.99290.0069 0.0001 4,221,813 29,544 494 2,983,209 1 0 0 2,983,209 0 0 7,205,022 29,544 494

1978 1,514,281 0.99290.0069 0.0001 1,503,583 10,522 176 3,687,577 1 0 0 3,687,577 0 0 5,191,160 10,522 176

1979 2,642,748 0.99290.0069 0.0001 2,624,077 18,363 307 4,990,113 1 0 0 4,990,113 0 0 7,614,190 18,363 307

1980 6,823,105 0.99290.0069 0.0001 6,774,901 47,411 793 5,292,727 1 0 0 5,292,727 0 0 12,067,628 47,411 793

1981 1,410,296 0.99290.0069 0.0001 1,400,332 9,800 164 1,687,147 1 0 0 1,687,147 0 0 3,087,479 9,800 164

1982 4,020,481 0.99290.0069 0.0001 3,992,077 27,937 467 5,380,891 1 0 0 5,380,891 0 0 9,372,969 27,937 467

1983 3,661,296 0.99290.0069 0.0001 3,635,430 25,441 425 5,380,327 1 0 0 5,380,327 0 0 9,015,756 25,441 425

1984 2,769,968 0.99290.0069 0.0001 2,750,399 19,247 322 5,561,882 1 0 0 5,561,882 0 0 8,312,281 19,247 322

1985 9,225,611 0.99290.0069 0.0001 9,160,435 64,105 1,072 4,618,524 1 0 0 4,618,524 0 0 13,778,959 64,105 1,072

1986 4,332,315 0.99290.0069 0.0001 4,301,708 30,103 503 3,922,368 1 0 0 3,922,368 0 0 8,224,076 30,103 503

1987 2,813,091 0.99290.0069 0.0001 2,793,218 19,547 327 3,927,949 1 0 0 3,927,949 0 0 6,721,167 19,547 327

1988 5,581,371 0.99290.0069 0.0001 5,541,940 38,782 648 6,159,420 1 0 0 6,159,420 0 0 11,701,360 38,782 648

1989 6,018,253 0.99290.0069 0.0001 5,975,736 41,818 699 6,996,223 1 0 0 6,996,223 0 0 12,971,959 41,818 699

1990 5,717,495 0.99290.0069 0.0001 5,677,102 39,728 664 4,663,560 1 0 0 4,663,560 0 0 10,340,662 39,728 664

1991 8,209,413 0.99290.0069 0.0001 8,151,416 57,044 954 4,526,781 1 0 0 4,526,781 0 0 12,678,197 57,044 954

1992 7,635,596 0.99290.0069 0.0001 7,581,653 53,056 887 227,467 1 0 0 227,467 0 0 7,809,120 53,056 887

1993 7,853,813 0.99290.0069 0.0001 7,798,328 54,573 912 414,276 1 0 0 414,276 0 0 8,212,604 54,573 912

1994 9,476,744 0.99290.0069 0.0001 9,409,793 65,850 1,101 402,445 1 0 0 402,445 0 0 9,812,239 65,850 1,101

1995 10,353,989 0.99290.0069 0.0001 10,280,841 71,945 1,203 535,573 1 0 0 535,573 0 0 10,816,414 71,945 1,203

1996 5,520,527 0.9929 0.0069 0.0001 5,481,526 38,360 641 0 1 0 0 0 0 0 5,481,526 38,360 641

66

Table.C5. Continued.

Inside Ocean Total

Age proportions Bycatch at age Age proportions Bycatch at age Bycatch at age

Year

DiscardedAtlantic croaker

bycatch(lbs.) Age0 Age1 Age2 Age0 Age1 Age2

Discarded Atlantic croaker

bycatch(lbs.)Age0Age1Age2 Age0Age1Age2 Age0 Age1 Age2

1997 8,882,533 0.99290.0069 0.0001 8,819,780 61,721 1,032 0 1 0 0 0 0 0 8,819,780 61,721 1,032

1998 3,645,992 0.99290.0069 0.0001 3,620,234 25,334 424 558,847 1 0 0 558,847 0 0 4,179,081 25,334 424

1999 3,166,786 1.00000.0000 0.0000 3,166,786 0 0 700,069 1 0 0 700,069 0 0 3,866,855 0 0

2000 4,712,732 1.00000.0000 0.0000 4,712,732 0 0 397,121 1 0 0 397,121 0 0 5,109,853 0 0

2001 2,096,044 1.00000.0000 0.0000 2,096,044 0 0 312,629 1 0 0 312,629 0 0 2,408,673 0 0

2002 4,510,500 1.00000.0000 0.0000 4,510,500 0 0 340,426 1 0 0 340,426 0 0 4,850,926 0 0

67

Figure C2. Annual landings of Atlantic croaker from North Carolina shrimp trawl fishery by general area, 1973-2002. Landings in 1996 and 1997 reflect a brief change in fishing behavior that was modified by regulation.

0

100,000

200,000

300,000

400,000

500,000

600,000

700,000

800,000

900,000

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

Year

Lan

din

gs (

lbs.

)

Ocean

Inside

Figure C1. Annual landings of Penaeid shrimp from North Carolina shrimp trawl fishery by general area, 1973-2002.

0

2,000,000

4,000,000

6,000,000

8,000,000

10,000,000

12,000,00019

73

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

Year

Lan

din

gs (

lbs.

)

Ocean

Inside

68

Figure C3. Annual discarded Atlantic croaker bycatch from the North Carolina shrimp trawl fishery, 1973-2002. Bycatch in weight (a) and number (b) are plotted for both the area pooled and alternative by area approaches.

a.

0

2,000,000

4,000,000

6,000,000

8,000,000

10,000,000

12,000,000

14,000,000

16,000,000

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

Cro

aker

Byc

atch

(L

bs.

)Lbs., Area Pooled

Lbs., By Area

b.

0

50,000,000

100,000,000

150,000,000

200,000,000

250,000,000

300,000,000

350,000,000

400,000,000

1973

1976

1979

1982

1985

1988

1991

1994

1997

2000

Year

Cro

aker

Byc

atch

(N

o.)

No., Area Pooled

No., By Area

69

Figure C4. Weighted length frequencies for NMFS and Johnson datasets by 10mm size intervals.

0

0.05

0.1

0.15

0.2

0.25

30 50 70 90 110

130

150

170

190

210

230

250

TL (mm)

Fre

qu

ency

NMFS

Johnson

Figure C5. Weighted length frequencies for NMFS Ocean and Inside waters considered separately, by 10mm size intervals.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

30 50 70 90 110

130

150

170

190

210

230

250

TL (mm)

Fre

qu

ency

Ocean

Inside

70

Figure C6. Atlantic croaker bycatch-at-age (No.) in the North Carolina shrimp trawl fishery, 1973-2002. Estimates are plotted for the area pooled and alternative by area approaches. Bycatch is plotted on log scale.

1

10

100

1,000

10,000

100,000

1,000,000

10,000,000

100,000,000

1,000,000,000

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

Year

Byc

atch

(N

o.)

Age 0 - area pooled

Age 1 - area pooled

Age 2 - area pooled

Age 0 - by area

Age 1 - by area

Age 2 - by area

71

APPENDIX D: An Evaluation of the NEFSC observer data to estimate Atlantic croaker discards “The Northeast Domestic Fisheries Observer Program collects, maintains and distributes data for scientific and management purposes in the northwest Atlantic Ocean (NEFSC Web Site). Since implementation in 1989, the Program has deployed an average of 35 observers a year in various commercial fisheries” (NEFSC Web Site). The Northeast Fisheries Science Center (NEFSC) observer data set contains information on Atlantic croaker discards from 1989 to 2002. This report provides a summary the available data and its potential use in estimating Atlantic croaker discards. Summary of the Data Set Between 1989 and 2002, a total of 1,267 observer-sampling trips caught or discarded Atlantic croaker. However, for 960 trips, discard records were not kept (G. Reppucci, NMFS NEFSC) and the data set was reduced to observations made on 306 trips. The majority of those sea days for the discarded trips were funded through Protected Species monies and data collection protocols required the observer to watch for mammal/turtle takes to ensure they don't miss 'fall outs' leading to incomplete discard data (G. Reppucci, NMFS NEFSC). Gill nets and fish otter trawls accounted for the majority of sampled trips (306 of 307 trips) and this analysis only evaluates these two gears.

For most years and both gears, discards were observed on less than 25 trips/year; gill nets accounted for 234 trips and fish otter trawls 72 trips over the entire time series (Table D1). The majority of observer sampling was carried out on vessels that landed their catch in Virginia and North Carolina (Table D2). In terms of target species, the majority of gill net trips targeted Atlantic croaker, with weakfish and spot being species of secondary species (Table D3). In general, kept landings of Atlantic croaker, weakfish and spot were likely to co-occur on the same trip. For otter trawls, the majority of sampled trips landed their catch in Maryland (Table D2) and targeted summer flounder (Table D3). Given the co-occurrence of the major target species and to avoid the possibility of “double counting” if estimates were based on species-specific targets, we developed discard estimates based on the landings of Atlantic croaker by gear type.

A comparison of the Atlantic croaker discarded to those kept by trip revealed weak relationships for the untransformed and transformed variables for both gears (Table D4). Linear regression on the log transformed variables revealed that the linear relationship between log (discards) and log (Landings) had R2 values of 0.12 and 0.03 for gillnets and otter trawls respectively (Figures D1 and D2). Linear regressions on discards to landings were also weak, with R2 values of 0.04 and 0.18 for gill nets and otter trawls respectively. Discard estimators were developed using a ratio approach and trip based approach. The ratio-based approach consisted of:

1) Developing discard ratios by gear type using all years of data. In addition, ratio estimators, by gear type and year were also estimated.

72

Discard ratios were estimated using Proc Survey means in SAS®, which uses the Taylor series expansion method to estimate the variance of the ratio estimator.

2) A regression approach where discards were estimated as a function of landings was also developed for each gear. Given the small sample size by year, regression estimators by year and gear were not developed. The linear regression estimators were based on log transformed discards and landings.

Trip based estimators were developed using two approaches: 1) Estimating the mean discards/trip and their variance by gear type and year,

and by gear type across all years. 2) Developing a general linear model (GLM) to estimate the mean discards

per trip. Mean discards per trip and their standard error by year were estimated using the least squares mean approach in SAS. The response variable was log discards and the explanatory variables were year, target species and number of hauls.

Estimates of total Atlantic croaker discards and their variance were determined for the gill net fishery and otter trawl fishery using a Monte-Carlo approach using the estimators discussed earlier. For the ratio-based estimators, annual estimates were based on the mean of 1,000 runs derived from the annual landings * an estimate based on a random value generated from the mean and standard error of the estimator (Table D5 and D6).

In order to develop a trip-based estimate, the annual number of trips by gear type was obtained from the Virginia and North Carolina trip ticket databases. For North Carolina, the annual number of gill nets and otter trawl trips were available from 1994-2002. For Virginia, only the numbers of gill net trips between 1993 and 2002 were available. As such, trip based estimates could only be developed for periods where trip data existed. Annual estimates of discards by state, gear and year were based on a Monte-Carlo approach where the mean of 1,000 runs were derived from the annual number of trips * a discard estimate based on a random value generated from the mean and standard error of the estimator (Table D7-D8).

Initial Observations

1) For the ratio-based method, the regression approach based on the log-log transformation produce very low estimates. While not presented here, we first looked at a regression model where the response variable log (discards+1) and the explanatory variable were landings. This was a poor model. The 1,000 Monte-Carlo runs produced estimates ranging from infinity to zero.

2) For the trip-based approach, we do not have effort information for the otter trawl-fishery for Virginia and it also assumes that the discard ratios observed in the coastal waters of the Atlantic ocean are applicable to the inshore gill net trips for Virginia. At best, trip based estimates can be estimated for the period 1993-2002, for all other periods an alternate approach would need to be used.

73

3) A ratio-based method would use a consistent methodology to estimates the entire time series, but the correlation between landings and discards is weak at best.

4) Estimates based on yearly samples by gear are poor, since the average number of trips sampled per year is low (< 25 trips).

74

Table D1. Summary of Atlantic croaker discard trips and associated hauls sampled in the NEFSC observer database for gill nets and otter trawls. Discards and landings in pounds. Gear Year Croaker Kept Croaker Discarded Hauls Trips Gill net 1993 2,701 95 56 8 Gill net 1994 53,730 824 162 28 Gill net 1995 71,915 403 211 24 Gill net 1996 121,313 1,012 248 28 Gill net 1997 111,052 318 167 20 Gill net 1998 91,871 77 207 36 Gill net 1999 14,557 55 82 20 Gill net 2000 50,520 205 156 30 Gill net 2001 41,436 667 115 22 Gill net 2002 23,958 24 92 18 Gill net Total 583,053 3,680 1496 234 Other 1995 0 31 9 1 Other Total 0 31 9 1 Trawl 1989 416 353 45 5 Trawl 1990 5 20 3 2 Trawl 1991 15 123 6 3 Trawl 1992 25 1,418 14 2 Trawl 1993 62 1,231 28 6 Trawl 1994 53,809 2,775 38 6 Trawl 1995 21,382 5,081 86 11 Trawl 1996 62,345 838 28 6 Trawl 1997 76,562 541 24 2 Trawl 1998 46,718 5,106 25 2 Trawl 1999 20,551 483 30 8 Trawl 2000 9,483 8 8 6 Trawl 2001 31,059 1,425 30 6 Trawl 2002 299 94 24 7 Trawl Total 322,731 19,496 389 72 Grand Total 905,784 23,206 1894 307

75

Table D2. Sampling trips of Atlantic croaker discards by location of state of landing.

Year MA MD NC NJ NY VA Grand Total

1989 0 4 0 1 0 0 5 1990 0 2 0 0 0 0 2 1991 0 2 0 0 0 1 3 1992 0 2 0 0 0 0 2 1993 0 6 2 0 0 6 14 1994 0 3 21 1 1 8 34 1995 1 5 11 5 1 13 36 1996 0 2 13 4 0 15 34 1997 0 0 7 1 0 14 22 1998 0 0 19 1 0 18 38 1999 0 0 12 1 1 14 28 2000 0 0 12 5 1 18 36 2001 0 3 9 4 0 12 28 2002 0 3 3 5 0 14 25

Grand Total 1 32 109 28 4 133 307

76

Table D3. Summary of sampling trips by target species is the species identified by fisher as his/her primary target on trip. Gill net Trawl Total Species Hauls Trips Hauls Trips Hauls Trips ALL COMBINED 2 1 0 0 2 1 BASS, STRIPED 27 7 0 0 27 7 BLUEFISH 24 6 0 0 24 6 BUTTERFISH 6 1 0 0 6 1 CRAB, HORSESHOE 0 0 31 4 31 4 CRAB,NK 0 0 7 1 7 1 CROAKER, ATLANTIC 763 98 74 11 837 109 DOGFISH, SPINY 46 8 0 0 46 8 DOGFISH,NK 2 1 0 0 2 1 FINFISH,NK 21 4 11 4 32 8 FLOUNDER, NK 2 1 0 0 2 1 FLOUNDER, SUMMER (FLUKE) 0 0 38 10 38 10 FLOUNDER,MIXED 0 0 6 1 6 1 FLOUNDER,SUMMER/FLUKE 0 0 185 33 185 33 GROUNDFISH,MIXED 7 1 1 1 8 2 KINGFISH, NK 9 3 0 0 9 3 MACKEREL, SPANISH 46 11 0 0 46 11 MENHADEN, ATLANTIC 7 1 0 0 7 1 MENHADEN/POGY 13 3 0 0 13 3 OTHER 14 3 8 1 22 4 SHARK,NK 4 1 0 0 4 1 SOUTHERN FLOUNDER 2 2 0 0 2 2 SPOT 178 23 0 0 178 23 WEAKFISH (SQUETEAGUE SEA TROUT) 23 6 0 0 23 6 WEAKFISH/BLUEFISH 9 2 0 0 9 2 WEAKFISH/CROAKER 31 5 20 1 51 6 WEAKFISH/GRAY SEA TROUT 260 46 8 5 268 51 Grand Total 1496 234 389 72 1885 306

77

Table D4. Correlation between Atlantic croaker discards and landings from the NEFSC Observer database.

Gill net Landings Discards Log (Discards+1) Log(Landings+1)Landings 1 Discards 0.201066 1 Log (Discards+1) 0.373964 0.616753 1 Log(Landings+1) 0.696137 0.205805 0.347876 1 Otter Trawl Landings Discards Log (Discards+1) Log(Landings+1)Landings 1 Discards 0.424636 1 Log (Discards+1) 0.246803 0.622602 1 Log(Landings+1) 0.654923 0.364162 0.185773 1

78

Table D5. Discard estimates of Atlantic croaker from the gill net fishery. Estimates in pounds. ratio_allyr= ratio estimator based on all years combined. reg_allyr= estimator based on log-log regression model. ratio_indyr=ratio estimator based on individual year. Gill net Total Discards Std Err

Year landingsratio_allyrreg_allyrratio_indyrratio_allyrreg_allyrratio_indyr1972 243,200 1,536 12 13 0 1973 619,300 3,848 14 31 0 1974 499,800 3,139 14 25 0 1975 802,700 5,039 16 42 0 1976 1,685,600 10,660 18 85 0 1977 2,934,600 18,419 20 144 0 1978 2,496,590 15,684 20 125 1 1979 2,363,480 15,010 20 122 1 1980 3,892,493 24,609 22 196 1 1981 1,369,646 8,622 18 71 0 1982 1,268,144 8,053 18 64 0 1983 924,657 5,881 17 48 0 1984 2,613,152 16,572 20 133 1 1985 2,821,276 17,816 20 144 0 1986 3,821,447 23,761 21 194 1 1987 3,037,713 19,072 20 153 0 1988 3,239,778 20,297 21 168 1 1989 1,784,728 11,204 18 90 0 1990 977,347 6,142 16 51 0 1991 975,819 6,213 17 50 0 1992 1,758,254 11,076 19 91 0 1993 3,816,993 23,645 20 130,650 191 0 1,5401994 4,672,895 29,472 23 71,643 243 1 6231995 4,917,110 31,561 24 28,344 254 1 3901996 8,130,019 51,373 26 68,041 410 1 1,3081997 8,056,062 50,171 25 22,540 428 1 331199810,713,447 67,935 28 8,980 557 1 1031999 7,837,131 49,919 27 30,482 413 1 790200010,610,535 66,685 27 42,967 539 1 330200111,236,025 70,130 26 173,953 551 1 4,8762002 9,548,509 60,894 27 9,718 487 1 272

79

Table D6. Discard estimates of Atlantic croaker from the otter trawl fishery. Estimates in pounds. ratio_allyr= ratio estimator based on all years combined. reg_allyr= estimator based on log-log regression model. ratio_indyr=ratio estimator based on individual year.

Trawl Total Discards Std Err Year Landings ratio_allyr reg_allyr ratio_indyrratio_allyr reg_allyrratio_indyr1972 3,225,400 197,458 322 2,395 37 1973 1,661,100 100,378 243 1,255 25 1974 2,522,900 151,248 261 1,898 26 1975 5,966,800 358,442 382 4,483 43 1976 10,872,000 655,587 376 7,941 34 1977 12,519,500 745,146 431 9,369 70 1978 12,948,089 787,861 506 9,431 96 1979 8,483,004 515,441 393 6,277 48 1980 5,615,381 333,831 269 3,945 27 1981 2,021,126 121,998 242 1,505 20 1982 1,859,491 114,232 279 1,338 40 1983 831,819 49,789 169 591 12 1984 2,439,191 145,237 235 1,774 21 1985 2,384,824 140,685 223 1,708 20 1986 2,500,141 151,267 275 1,808 29 1987 2,121,088 125,756 233 1,604 20 1988 2,017,540 125,205 277 1,464 25 1989 1,638,547 98,918 220 1,389,308 1,196 17 20,7001990 343,752 20,943 159 1,411,653 258 11 53,3201991 339,505 20,454 134 2,775,374 245 8 38,6631992 828,181 49,346 217 46,656,967 641 22 298,1451993 2,215,984 134,804 306 44,260,991 1,695 33 473,5211994 3,225,941 195,164 304 166,356 2,417 36 1661995 4,305,209 261,321 349 1,035,900 3,135 48 32,2141996 5,781,415 350,658 378 78,779 4,335 49 3,2771997 10,231,066 618,572 351 72,321 7,409 36 3691998 6,809,853 412,814 419 745,530 5,332 46 4,6251999 9,562,721 584,455 426 232,899 7,182 48 8,5212000 8,913,884 538,543 377 7,523 6,753 37 3112001 8,615,978 521,200 379 397,002 6,357 45 14,8432002 7,990,650 479,083 329 2,470,228 5,894 28 68,263

80

Table D7. Estimated Atlantic croaker discards in the gill net fishery using trip based estimators. trip_lsm_va= discard estimates for Virginia using the annual least squares estimates. trip_lsm_nc = discard estimates for North Carolina using the annual least squares estimates. trip_ind_va= discard estimates for Virginia based on the average annual discard weight per trip. trip_ind_nc= discard estimates for North Carolina based on the average annual discard weight per trip. trip_all_va= discard estimates for Virginia based on the average discards per trip based across all years sampled. trip_all_nc= discard estimates for North Carolina based on the average discards per trip based across all years sampled. All estimates are in pounds.

Gill net

Year Landingstrip_lsm

_vatrip_lsm

_nctrip_ind_

vatrip_ind

_nctrip_all_

vatrip_all_

nctrip_lsm_va

trip_lsm_nc

trip_ind_va

trip_ind_nc

trip_all_va

trip_all_nc

1993 3,816,993 63,937 110,431 147,666 971 1,359 1,191 1994 4,672,895 104,765 107,400 313,262 321,143 167,409 171,621 883 905 4,503 4,617 1,381 1,4161995 4,917,110 39,047 57,025 172,312 251,647 159,788 233,357 348 508 2,253 3,291 1,288 1,8801996 8,130,019 27,356 41,013 312,144 467,981 135,545 203,216 224 336 5,881 8,817 1,083 1,6241997 8,056,062 39,950 41,635 190,967 199,021 190,211 198,233 395 412 2,458 2,562 1,624 1,6931998 10,713,447 22,169 18,083 24,491 19,977 180,973 147,619 162 132 264 215 1,483 1,2101999 7,837,131 19,437 23,311 26,166 31,382 146,727 175,978 196 235 673 807 1,215 1,4572000 10,610,535 32,437 30,346 74,135 69,355 170,503 159,510 254 238 787 736 1,380 1,2912001 11,236,025 26,911 31,439 270,533 316,046 144,025 168,255 237 277 7,223 8,438 1,133 1,3242002 9,548,509 15,538 11,537 13,985 10,384 164,832 122,388 159 118 342 254 1,318 979

Discard Estimates Standard Error(Trip based in Pounds)

81

Table D8. Estimated Atlantic croaker discards in the otter trawl fishery using trip based estimators. trip_lsm_va= discard estimates for Virginia using the annual least squares estimates. trip_lsm_nc = discard estimates for North Carolina using the annual least squares estimates. trip_ind_va= discard estimates for Virginia based on the average annual discard weight per trip. trip_ind_nc= discard estimates for North Carolina based on the average annual discard weight per trip. trip_all_va= discard estimates for Virginia based on the average discards per trip based across all years sampled. trip_all_nc= discard estimates for North Carolina based on the average discards per trip based across all years sampled. All estimates are in pounds.

Year Landingstrip_lsm_va

trip_lsm_nc

trip_ind_va trip_ind_nc

trip_all_va trip_all_nc

trip_lsm_va

trip_lsm_nc

trip_ind_va

trip_ind_nc

trip_all_va

trip_all_nc

1994 3,225,941 12,684 104,901 61,283 552 3,274 7321995 4,305,209 5,920 112,019 65,287 170 2,917 7561996 5,781,415 6,016 40,062 77,200 237 1,215 9211997 10,231,066 17,393 108,172 108,128 1,659 3,364 1,2501998 6,809,853 170,390 1,507,144 158,938 16,986 50,317 1,9811999 9,562,721 24,463 24,300 108,168 994 467 1,2832000 8,913,884 4,342 405 82,325 195 9 9962001 8,615,978 2,368 61,915 70,495 91 1,446 8302002 7,990,650 1,303 3,065 61,827 50 38 734

Discard Estimates (Trip Based in Pounds) Standard Error

82

Figure D1. Linear relationship between the log transformed discards and log transformed landings for gill net samples.

Figure D2. Linear relationship between the log transformed discards and log transformed landings for otter trawls samples.

y = 0.1822x + 0.1305R2 = 0.121

0

1

2

3

4

5

6

7

0 2 4 6 8 10 12

Log Landings (pounds)

Log

Dis

card

s (p

ound

s)

y = 0.135x + 2.161R2 = 0.0345

0

1

2

3

4

5

6

7

8

9

0 2 4 6 8 10 12

Landings (pounds)

Dis

card

s (p

ound

s)

83

Figure D3. Linear relationship between the discards and landings for gill net samples. Figure D4. Linear relationship between the discards and landings for otter trawls samples

y = 0.0276x + 146.95R2 = 0.1803

0

1000

2000

3000

4000

5000

6000

0 10000 20000 30000 40000 50000 60000 70000 80000

Landings (pounds)

Dis

card

s (p

ound

s)

y = 0.0032x + 7.8R2 = 0.0404

0

100

200

300

400

500

600

700

0 5000 10000 15000 20000 25000

Landings (pounds)

Dis

card

s (p

ound

s)

84

APPENDIX E: Using the NMFS-Woods Hole Fall Groundfish Survey off the East Coast of the United States: 1972-2002 to develop indices of abundance for Atlantic croaker The Woods Hole laboratory of the National Marine Fisheries Service has been conducting stratified random surveys of fishes from Long Island to Cape Hatteras since 1982. The division of the region into a series of depth zones (0 – 9 m; 9 – 18 m; 18 – 27 m; 27 – 55 m; 55 – 110 m; 110 – 188 m; 188 – 366 m) provides the framework for the erection of strata based on latitude and depth. The area within each stratum is subdivided into one-nautical mile blocks that are selected randomly prior to the sampling trip (see Azarovitz, 1994 for a review of the survey; see also Grosslein, 1969). At each randomly selected station, a trawl tow was made according to a standardized format. A #36-Yankee otter trawl rigged with rollers, 5 fathom legs and 1000 pound polyvalent door was fished for 30-minutes. The cod end and upper belly were lined with 0.5-inch mesh to retain young-of-year fish. The data recorded at each station (excluding the catch itself) are in Table E1. Variable Columns Definition Cruise code 1-4 First two digits are the year; second two digits sequentially

assigned to indicate the order in which the cruise were coded Station 5-8 Sequential station number Stratum 9-13 Strata identification numbers Tow number 14-16 Number of tow within a given stratum Station value 17 Station type – [1] survey haul; [2] non-random haul; [3]

special random add-on station haul [4] comparison haul; [5] no trawl – other gear; [6] site=specific; [7] systematic grid; [8] depletion site; [9] systematic parallel tracts; [0] systematic zigzag transects.

18 Relative success of haul [1] good tow – no gear or tow duration problems; [2] representative but with some problems due to gear or tow time; [3] problem tow – may or may not be representative due to gear or tow time; [4] not representative due to gear or time; [5] no trawl tow – other gear used

19 Gear condition – [1] no damage to insignificant damage; [2] twisted wing small tears in belly; [3] hang with no to minor damage; [4] parted legs, sweep or headrope; [5] tear-up. But not total; [6] obstruction in net; [7] crossed doors; [8] open gear; [9] hang with major damage

Statistical area 20-22 International statistical area where tow made Vessel 23-24 AL = Albatross IV; AT = Atlantic Twin; DE = Delaware II Cruise 25-26 Assigned vessel cruise number Time 27 [1] = Eastern standard time; [2] = eastern daylight savings

time Yr-mo-day 28-33 First two = year; second two = month; last two = day Gear code 34-35 Type of gear used (numerous code values) Time 36-39 Local time with 24-hour clock Minutes out 40-42 Actual time for gear out to with tenths of minutes Depth start/end 43-50 First four digits = start depth; last four digits = end depth Min/max 51-58 First four digits = minimum depth; second four digits =

maximum depth

Table E1. Data recorded at each station during a groundfish survey cruise.

85

Table E1. Continued. Variable Columns Definition Lat/long 59-66 Replace by GPS reading in 1999; prior to that, first four digits

= beginning latitude last four digits beginning longitude; both rounded to whole minutes

Loran 67-98 Loran readings Cable 99-102 Cable in water measured in meters at the water’s surface Pitch 103-105 Pitch of prop if applicable Heading 106-108 Vessel’s heading in degrees Course 109-111 Actual course the vessel made good in degrees Rpm 112-114 Average shaft rpm under tow Doppler bottom 115-117 Speed over bottom Doppler water 118-120 Used on special occasions Des speed 121-122 Designated towing speed for a particular gear to 0.1 knots Gear id 123-124 Gear id number (each net has it’s own number) Door id 125-126 Door id number Head-rope height 127-129 Not used leave blank Other gear 130-131 Code for other gear (hydro, plankton, etc) Air temp 134-136 Air temp to nearest whole degree Cloud cover 137-138 Code for cloud cover Insol 139-142 Not used – blank Bar 143-146 Barometric pressure to nearest millibar Wind dir & speed 147-151 Wind direction in degrees; wind velocity in knots Weather 152-153 Weather codes Wave height 154-155 Height of waves to nearest tenth of meter Swell direction & hgt 156-160 First three digits swell direction in degrees, second two digits

= swell height in tenths of meters Surf temp 161-163 Surface temp in tenths of degrees C Surf salinity 164-167 Salinity parts per 1000 Wingspread 168-171 Not used – blank Sal depth 172-175 Not used leave blank Xbt 176 Type of temperature profiler used Surf & bot temp 177-182 First three digits = surface degrees c to tenths; second three

digits = bottom from xbt or ctd Coded species 183-184 Number of species caught and coded at station Trash 185-188 Amount of trash in liters Fullness of dredge 189-191 Mot used leave blank Sed type 192-194 Not used leave blank Trash by % 195-203 Not used leave blank Ave depth 203-207 Average depth in meters between start and end of tow Calc speed 208-210 Calculated speed of tow derived from navigational

instruments Surf sal 214-218 Surface salinity (0.001) Bot sal 219-223 Bottom salinity (0.001) Total weight 224-229 Total weight of species to 0.1 kg Total number 230-235 Total number of animals at station Trawl surveys are used to generate abundance estimates, distributional patterns, hydrographic data and specimens for life history studies do not have the experimental rigidity of laboratory studies. For example, severe weather, mechanical problems with the vessel, and at sea illnesses are just a few of the difficulties that are encountered during a cruise. There are a considerable number of strata in the Mid-Atlantic area surveyed over the past 20+ years. Periods of equipment failure as well as poor weather resulted in

86

some strata not being sampled with the same intensity through time. Indeed, in the early years, the cruise only sampled a single station in the shallowest depth stratum. Other depth zones on other occasions had a limited number of tows so that the within stratum variance either could not be calculated or was based on a few observations. The problems of insufficient sampling intensity within various strata are not new. When summarizing the Marine Resources Monitoring Assessment and Prediction (MARMAP) trawl survey data for the South Atlantic Bight (Cape Hatteras, NC to Cape Canaveral, FL), examination of the distribution of the tows within a stratum during a particular survey indicated that it was unrealistic to calculate stratum means and variances with existing sample sizes. After considering alternatives, it was decided that the strata could be collapsed within a depth zone. This resulted in a reasonable number of tows available for estimates of the mean catch per standard tow and its variance within a depth zone. Since the areas of each stratum was known, a stratified mean catch per tow with its associated standard error could be calculated for the number and weight of the catch (see Wenner et al. 1979). In addition to within stratum sample size, additional problems are frequently encountered in trawl survey data. What constitutes an acceptable tow? When should a tow be eliminated from the data set because of problems? The NMFS groundfish survey of the eastern Atlantic coast has established a series of coded observations that include the time a net is towed at a station in comparison to the standard tow time of 30-minutes, and the condition of the trawl net at the end of the haul (see Table E1). This is recorded as the variable “SHG” in the data set. The inclusion of a specific tow is a judgment made by investigators during the analysis of the data. Some have not included tows that had an shg greater than 136 [= a stratified random trawl tow (first digit =1); that may or may not be representative of the site due to tow time as short as 20-25 minutes or a long as 35-40 minutes (second digit = 3) and the condition of the gear (third digit less than 6). Method for the present analysis Station, catch, and length frequency data for Atlantic croaker were obtained for the Groundfish Survey Unit of the National Marine Fisheries Service Science Center at Woods Hole, MA. The area requested were trawl tows made between southern Long Island, NY south to Cape Hatteras, NC in the three shallowest depth zones (0 – 9m; 9 – 18m; 18 – 27m) for the fall survey (generally completed in September) from 1982 to 2002.

87

These data were edited for appropriateness for inclusion in the analysis. The criteria used were: Inclusion criteria Rationale Tow made in depths less than 27 m Atlantic croaker are extremely rare in samples

deeper than 27 m along the east coast of the US; inclusion of deeper tows would provide no additional information

Tows made in strata from 3180 to 3440 Atlantic croaker were taken in only 3.8% of the 343 tows made in adjacent strata to the North (Table 2)

Tows that had an shg value (= tow conditions) less than 125

Based on the Woods Hole code, it was felt that the inclusion of only these tows would be a conservative approach

The use of more stringent criteria for tow condition resulted in the elimination of 33 tows from the original data set for the included strata. This was in contrast to the exclusion of 10 tows with the more liberal shg value of 136. As previously mentioned, not all strata were sampled each year and when sampled, often the intensity was low. Two approaches could be taken in the calculation of an index of relative abundance from these data. First, those strata that were not sampled consistently through the time series could be eliminated from the data set. Second, strata within depth zones could be collapsed so that each zone could be treated as a large stratum with trawl sites occupied throughout

88

Stratum

Old # New #

Mean latitude

Tows with Atlantic croaker

Total number of tows 12 3120 40.6 2 34 13 3130 40.5 1 70 14 3140 40.5 1 81 15 3150 40.0 2 21 16 3160 40.0 5 69 17 3170 40.0 2 68 18 3180 39.5 6 27 19 3190 39.5 12 70 20 3200 39.5 7 69 21 3210 39.1 5 25 22 3220 39.1 18 64 23 3230 39.1 9 63 24 3240 38.9 22 44 25 3250 38.8 38 67 26 3260 38.8 31 63 27 3270 38.3 18 24 28 3280 38.3 50 69 29 3290 38.4 37 64 30 3300 38.0 25 30 31 3310 37.9 60 69 32 3320 37.9 40 62 33 3330 37.5 26 27 34 3340 37.4 51 67 35 3350 37.4 44 63 36 3360 37.0 34 44 37 3370 37.0 43 70 38 3380 36.9 43 66 39 3390 36.3 28 31 40 3400 36.3 58 65 41 3410 36.3 49 69 42 3420 35.7 24 29 43 3430 35.6 54 64 44 3440 35.7 45 67 45 3450 41.4 Out of area 46 3460 41.1 Out of area 52 3520 41.0 Out of area 55 3550 41.2 Out of area

The analysis of the frequency of tows made in strata by year indicated that the shallowest strata (0 – 9m) would be eliminated because of either missing years or a year with a single tow. In 1987 and 1988, no tows were made in any of the shallow strata. Over the time series, Atlantic croaker were taken in 67% of the 274 tows made in 0 – 9m, 64% of those in 9 – 18m (n = 595) and 52% of the 570 tows in the deepest zone retained in the data set. The overall catch per tow of Atlantic croaker in numbers was also higher (404) in the 0 – 9m depth zone than those in the 9 –18m (262) or 18 – 27m (153) zones.

Table E2. Presence-absence of Atlantic croaker in strata from near New York south to Cape Hatteras with the average latitude for all tows in that stratum. Old # refers to historical identification number of strata by NMFS-Woods Hole; New # refers to present identification number of strata.

89

During each of the surveys, most strata had only two tows (Table E2). This would result in estimates of the within stratum variances being very large and having only a single degree of freedom. For this reason, I determined that the best approach would be to collapse the strata within a given depth zone into a large stratum that would include the heart of the species distribution along the Middle Atlantic Coast. The resultant distribution of stations within the depth zones is in Table E3. The locations of the various strata are in Figure E1. The stratified mean catch per tow for the three depth “strata” were calculated for each cruise using the following formula (Krebs,1989):

NxNyL

hhhst /

1

where sty = the stratified mean catch per tow

hN = size of stratum h

hx = mean catch per tow for the h stratum

N = total population size = hN

The variance for each of the three strata (= collapsed depth zones) was determined with the equation

L

hhhhhst fnswyiance

1

22 1/*var

where hw = stratum weight = Nh/N

2hs = observed variance of stratum h

nh = number of tows in stratum h fh = sampling fraction in stratum h = nh/Nh The standard error of the stratified mean is

standard error of sty = styiancevar

90

Stratum

Year 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

3180 2 3 0 2 1 1 1 1 1 1 0 0 1 1 1 0 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 3190 4 3 2 3 2 2 2 2 2 2 3 2 2 2 2 2 2 3 3 2 2 2 2 2 2 2 2 2 2 2 2 3200 3 4 4 2 2 2 2 2 2 2 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 3210 3 2 0 2 1 0 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 0 1 0 0 0 0 0 1 1 3220 3 2 3 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3230 3 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3240 3 2 0 2 1 1 1 1 1 1 1 1 1 1 1 0 0 2 2 2 1 2 1 2 2 2 2 1 2 2 2 3250 4 3 3 2 2 2 2 2 2 2 0 2 2 2 2 2 2 2 3 2 2 2 2 2 1 2 2 2 2 2 2 3260 3 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 3270 1 2 0 2 1 1 1 1 1 1 0 0 1 1 0 0 0 1 0 1 1 1 1 1 1 0 1 1 0 0 1 3280 3 3 3 2 2 2 2 2 2 2 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 2 3290 3 2 3 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 3300 4 2 0 2 1 1 1 0 1 1 0 0 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3310 3 4 3 3 2 2 2 2 2 2 3 3 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3320 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3330 2 1 0 2 1 1 1 1 1 1 0 1 1 1 1 0 0 1 1 1 1 1 1 1 0 1 1 1 0 1 1 3340 3 3 3 2 2 2 2 2 2 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 3 2 2 3350 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 3360 3 2 2 1 1 1 1 1 1 1 1 1 1 1 0 0 0 1 2 2 1 2 2 2 1 2 2 2 2 2 2 3370 4 4 4 3 2 2 2 2 2 2 2 2 2 2 2 2 2 3 2 2 1 2 2 2 2 2 2 2 2 2 2 3380 3 3 4 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3390 2 2 0 2 1 1 1 1 1 1 0 1 1 1 2 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3400 3 3 2 2 2 2 2 2 2 2 3 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 3410 4 4 4 2 2 2 2 2 2 2 0 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 3420 2 2 0 1 1 1 1 1 1 1 0 1 1 1 1 0 0 0 1 1 1 1 1 1 0 1 1 0 1 1 1 3430 3 2 0 4 2 2 2 2 2 2 1 2 2 2 2 2 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 3440 2 4 4 3 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Total 77 70 52 58 45 44 45 44 45 45 35 45 45 44 42 35 36 47 46 47 43 47 44 47 41 44 45 43 45 46 47

Table E3. Number of tows made in each of the strata include in the analysis for each year’s survey.

91

Depth Zone Year 0 – 9 m 9 – 18 m 18 – 27 m 1972 22 30 25 1973 18 27 25 1974 2 23 27 1975 16 23 19 1976 9 18 18 1977 8 18 18 1978 9 18 18 1979 8 18 18 1980 9 18 18 1981 9 18 18 1982 3 13 19 1983 6 20 19 1984 9 18 18 1985 9 18 17 1986 8 17 17 1987 0 18 17 1988 0 18 18 1989 8 21 18 1990 9 20 17 1991 11 18 18 1992 9 16 18 1993 11 18 18 1994 9 18 17 1995 11 18 18 1996 7 17 17 1997 9 17 18 1998 10 18 17 1999 7 18 18 2000 7 20 18 2001 10 19 17 2002 11 18 18

Table E4. The number of stations occupied in each of the three shallow depth

92

Figure E1. Chart of strata mentioned or used in this report. The older stratum numbers are used in order to save space. See Table E2 for the recent designations of these strata. Strata not drawn exactly as indicated on nautical charts because of space considerations.

93

Results Means, variances, and weighting factors for strata collapsed into depth zones for each year are in Table E5. The stratified mean catch per tow with the associated standard error for each year of the time series is in Table E6.

Figure E2. The stratified mean catch per tow +/- one standard error of the mean for the number of Atlantic croaker caught during the southern leg of the fall groundfish survey conducted by NMFS-Woods Hole. Strata included are from mid-New Jersey to Cape Hatteras (see figure E1). These were collapsed into three depth zones. Solid horizontal line is the mean of the time series. During the earlier period of the time series, the stratified mean catch per tow peaked in 1975 and then fell to some of the lowest values observed during the early 1980’s. A slight increase in the catches in relation to the long-term mean was followed by another decline in the mid-1980’s. Since then, the catches of Atlantic croaker in the groundfish survey have shown variability between years with a general upward trend. The two highest estimates have been during survey made during the last four years.

94

Year

Zone

Mean number

yst

Variance number

sh2

Number of tows nh

Number tows with Atlantic croaker

Area (nautical

miles2)Nh

Area surveyed

Nh

Stratum weight wh Nh /N

wh * yh

wh2 * sh

2 /nh *1 fh

1972 0-9 9.82 778.7 22 12 568 4423 0.128 1.26 0.56 1972 9–18 2.17 50.1 30 5 1888 4423 0.427 0.92 0.30 1972 18-27 0.32 1.5 25 3 1967 4423 0.445 0.14 0.01 1973 0-9 48.7 5263.6 18 11 568 4423 0.128 6.25 4.67 1973 9–18 50.7 5226.6 27 19 1888 4423 0.427 21.64 34.77 1973 18-27 22.9 1833.8 25 10 1967 4423 0.445 10.18 14.32 1974 0-9 446.5 171112.5 2 2 119 3802 0.031 13.98 82.41 1974 9–18 232.9 454037.0 23 15 1716 3802 0.451 105.12 3967.47 1974 18-27 46.6 7530.9 27 13 1967 3802 0.517 24.11 73.63 1975 0-9 2076.8 13544942.4 16 10 568 4423 0.128 266.70 13567.84 1975 9–18 767.4 1915637.4 23 16 1888 4423 0.427 327.57 14991.06 1975 18-27 98.8 50510.5 19 8 1967 4423 0.445 43.93 520.70 1976 0-9 583.8 390166.4 9 7 568 4423 0.128 74.97 703.61 1976 9–18 625.2 903084.6 18 13 1888 4423 0.427 266.88 9054.53 1976 18-27 125.4 182967.1 18 9 1967 4423 0.445 55.76 1991.97 1977 0-9 209.5 79438.6 8 6 546 4401 0.124 25.99 150.60 1977 9–18 141.3 128091.9 18 13 1888 4401 0.429 60.61 1297.15 1977 18-27 73.3 26946.9 18 10 1967 4401 0.447 32.75 296.31 1978 0-9 729.3 4251475.2 9 7 568 4423 0.128 93.66 7666.96 1978 9–18 105.5 32588.6 18 13 1888 4423 0.427 45.03 326.74 1978 18-27 51.8 30912.4 18 11 1967 4423 0.445 23.03 336.54 1979 0-9 26.7 1628.2 8 5 493 4348 0.113 3.03 2.57 1979 9–18 21.2 4222.7 18 10 1888 4348 0.434 9.19 43.81 1979 18-27 7.6 288.4 18 10 1967 4348 0.452 3.42 3.25 1980 0-9 50.0 5808.0 9 4 568 4423 0.128 6.42 10.47 1980 9–18 96.3 64416.9 18 6 1888 4423 0.427 41.12 645.86 1980 18-27 92.2 131456.3 18 8 1967 4423 0.445 40.99 1431.17 1981 0-9 9.0 670.5 9 2 568 4423 0.128 1.16 1.21 1981 9–18 4.1 85.5 18 10 1888 4423 0.427 1.73 0.86

Table E5. Mean catch per tow, variance and number of tows by depth zone and year.

95

Table E5. Continued. Year

Zone

Mean number

yst

Variance number

sh2

Number of tows nh

Number tows with Atlantic croaker

Area (nautical

miles2)Nh

Area surveyed

Nh

Stratum weight wh Nh /N

wh * yh

wh2 * sh

2 /nh *1 fh

1981 18-27 64.9 34372.5 18 10 1967 4423 0.445 28.88 374.22 1982 0-9 0.3 0.3 3 1 194 3494 0.056 0.019 0.0 1982 9–18 5.5 87.8 19 10 1716 3494 0.491 2.71 1.10 1982 18-27 14.1 1784.2 13 3 1584 3494 0.453 6.38 27.98 1983 0-9 690.0 1033048.4 6 5 361 4216 0.086 59.08 1241.38 1983 9–18 355.3 1334961.8 20 9 1888 4216 0.448 159.11 13243.92 1983 18-27 29.5 5333.2 19 9 1967 4216 0.467 13.75 60.51 1984 0-9 278.1 389360.8 9 6 568 4423 0.128 35.71 702.16 1984 9–18 356.8 615286.3 18 14 1888 4423 0.427 152.29 61.69.00 1984 18-27 179.0 91768.7 18 13 1967 4423 0.445 79.61 991.09 1985 0-9 72.7 8489.3 9 7 568 4423 0.128 9.33 15.31 1985 9–18 338.5 555703.0 18 13 1888 4423 0.427 144.49 5571.60 1985 18-27 135.2 91759.6 17 12 1967 4423 0.445 60.14 1058.30 1986 0-9 84.7 17395.1 8 5 414 4269 0.097 8.22 20.05 1986 9–18 137.0 233906.3 17 12 1888 4269 0.442 60.59 2666.96 1986 18-27 126.5 80048.5 17 11 1967 4269 0.461 58.30 991.04 1987 9–18 57.5 16121.1 18 10 1888 3855 0.490 28.16 212.77 1987 18-27 164.2 414321.1 17 8 1967 3855 0.510 83.80 6290.40 1988 9–18 50.6 25459.2 18 10 1888 3855 0.490 24.76 366.02 1988 18-27 13.5 581.8 18 9 1967 3855 0.510 6.89 8.34 1989 0-9 41.0 5156.6 8 3 431 4286 0.101 4.12 6.40 1989 9–18 43.7 8852.0 21 14 1888 4286 0.441 19.26 80.88 1989 18-27 166.2 222314.1 18 11 1967 4286 0.458 76.26 2577.55 1990 0-9 168.6 191049.0 9 5 436 4291 0.102 17.13 214.63 1990 9–18 58.1 14697.5 20 9 1888 4291 0.440 25.56 140.76 1990 18-27 81.0 38598.5 17 9 1967 4291 0.458 37.13 472.98 1991 0-9 677.0 2984745.6 11 7 568 4423 0.128 86.94 4388.18 1991 9–18 403.7 797433.9 18 12 1888 4423 0.427 172.33 7995.25 1991 18-27 2.8 68.1 18 5 1967 4423 0.445 1.26 0.74

96

Year Zone Mean Number

Variance Number

Number of Tows

Number tows with Atlantic

croaker

Area (nautical

miles2

Area surveyed

Stratum Weight

wh * yh wh2 * sh

2 /nh *1 fh

1992 9–18 354.8 399059.0 16 12 1888 4423 0.427 151.46 4506.00 1992 18-27 22.2 4601.5 18 8 1967 4423 0.445 9.88 50.10 1993 0-9 940.7 5912807.0 11 7 568 4423 0.128 120.81 8693.02 1993 9–18 41.1 9944.5 18 9 1888 4423 0.427 17.52 99.71 1993 18-27 5.7 397.6 18 3 1967 4423 0.445 2.54 4.33 1994 0-9 239.9 224953.1 9 6 546 4401 0.124 29.76 336.02 1994 9–18 153.3 78652.8 18 10 1888 4401 0.429 65.76 796.49 1994 18-27 857.1 2286676.9 17 9 1967 4401 0.447 383.06 26637.42 1995 0-9 245.7 380558.6 11 7 568 4423 0.128 31.56 559.50 1995 9–18 210.2 120550.0 18 14 1888 4423 0.427 89.71 1208.66 1995 18-27 153.1 45159.9 18 16 1967 4423 0.445 68.09 491.66 1996 0-9 305.6 288085.6 7 5 414 4269 0.097 29.63 380.51 1996 9–18 161.0 157272.7 17 13 1888 4269 0.442 71.20 1793.20 1996 18-27 223.9 137513.4 17 12 1967 4269 0.461 103.16 1702.48 1997 0-9 414.7 566242.7 9 6 511 4366 0.117 48.53 846.68 1997 9–18 159.4 213930.1 17 13 1888 4366 0.432 68.91 2332.01 1997 18-27 92.6 60556.4 18 9 1967 4366 0.451 41.70 676.61 1998 0-9 395.1 242933.4 10 9 546 4401 0.124 49.02 367.06 1998 9–18 476.1 638218.6 18 16 1888 4401 0.429 204.22 6463.05 1998 18-27 204.8 161720.1 17 13 1967 4401 0.447 91.54 1883.87 1999 0-9 451.3 349764.2 7 6 409 4264 0.096 43.29 451.85 1999 9–18 651.5 916928.7 18 16 1888 4264 0.443 288.47 9891.73 1999 18-27 872.9 2694278.1 18 15 1967 4264 0.461 402.69 31561.07 2000 0-9 662.0 1204226.6 7 7 322 4177 0.077 51.03 1000.11 2000 9–18 136.0 36312.7 20 15 1888 4177 0.452 61.47 367.01 2000 18-27 584.3 1882474.3 18 12 1967 4177 0.471 275.14 22979.67 2001 0-9 110.1 31728.5 10 8 533 4388 0.121 13.37 45.94 2001 9–18 327.8 657566.5 19 14 1888 4388 0.430 141.06 6342.55 2001 18-27 51.8 18676.4 17 8 1967 4388 0.448 23.20 218.85 2002 0-9 1314.1 8270889.2 11 10 568 4423 0.128 168.76 12159.88 2002 9–18 1606.3 19006251.7 18 16 1888 4423 0.427 685.66 190560.90 2002 18-27 192.1 136963.3 18 11 1967 4423 0.445 85.41 1491.13

97

Year Number of Tows Number of Tows

with Atlantic croaker

Stratified mean Number /tow

Standard error stratified mean

number/tow 1972 77 20 2.33 0.93 1973 70 40 38.07 7.33 1974 52 30 143.20 64.21 1975 58 34 638.21 170.53 1976 45 29 397.61 108.40 1977 44 29 119.35 41.76 1978 45 31 161.72 91.27 1979 44 25 15.64 7.05 1980 45 18 88.53 45.69 1981 45 22 31.77 19.40 1982 35 14 9.11 5.39 1983 45 23 231.94 120.61 1984 45 33 267.61 88.71 1985 44 32 213.97 81.52 1986 42 28 127.11 60.65 1987 35 18 111.96 80.64 1988 36 19 31.65 18.56 1989 47 28 99.64 51.62 1990 46 23 79.82 28.78 1991 47 24 260.53 111.28 1992 43 27 216.19 74.90 1993 47 19 140.88 93.79 1994 44 25 478.57 166.77 1995 47 37 189.36 47.54 1996 41 30 203.99 62.26 1997 44 28 159.14 62.09 1998 45 38 344.79 93.35 1999 43 37 734.45 204.71 2000 45 34 387.65 156.03 2001 46 30 177.64 81.29 2002 47 37 939.82 451.90

After looking at the index of abundance (stratified mean catch per tow), the data was examined to determine, within the core strata used in the analysis, where did Atlantic croaker occur more frequently, where was the greatest abundance over the time series, and were there any spatial differences in size of the fishes caught during the survey. The catch data was pooled, by both number per tow and number of occurrences, as well as the length frequencies over the time series by stratum. The stratum used were the Woods Hole designations of strata, not the depth zones used to calculate the stratified means. The highest mean catches for the 1972-2002 period were in strata 3390 (39) where tows yielded over 800 Atlantic croakers per haul (Table E7). In general, the largest catches were in the lower latitudes and in the shallower strata. The frequency of occurrence also increased with decreasing latitude (Table E8).

Table E6. Stratified mean catch per tow in number of Atlantic croaker with frequency of occurrence in samples and the standard error of the stratified mean.

98

Stratum N MeanStd.

Error Minimum Maximum

3180 26 379.42 370.66 0 96443190 69 49 34.071 0 22653200 67 2.03 1.808 0 1213210 25 127.76 106.153 0 26173220 63 27.97 13.513 0 7663230 62 14.23 9.337 0 5623240 43 137.88 54.942 0 18723250 64 119.69 66.785 0 40703260 61 62.05 21.017 0 9313270 23 433.52 138.246 0 22683280 69 186.26 48.987 0 25083290 62 127.65 40.413 0 14043300 30 368.57 133.273 0 33233310 68 250.1 64.473 0 26433320 62 227.02 87.371 0 44293330 27 481.93 230.442 0 57583340 66 618.27 294.656 0 186163350 60 354.13 134.229 0 63013360 43 595.86 317.35 0 133063370 69 320.43 104.254 0 51613380 65 85 40.192 0 24673390 31 849.45 370.736 0 83803400 64 355.17 84.332 0 32153410 65 313.32 108.876 0 50743420 26 537.04 246.343 0 62253430 63 379.76 109.415 0 54593440 66 158.71 43.931 0 2071Total 1439 247.11 24.93 0 18616

Table E7. Mean number per tow with appropriate statistic for Atlantic croaker taken in the southern leg of the fall groundfish survey in core strata. Data pooled over all years to derive values.

99

Southern Leg Fall Groundfish Survey - NMFS

Atlantic Croaker

Stratum Presence-Absence

Total tows Absent in tow Present in tow

3180 21 5 26 3190 57 12 69 3200 61 6 67 3210 20 5 25 3220 45 18 63 3230 53 9 62 3240 22 21 43 3250 27 37 64 3260 30 31 61 3270 5 18 23 3280 19 50 69 3290 26 36 62 3300 5 25 30 3310 9 59 68 3320 22 40 62 3330 1 26 27 3340 15 51 66 3350 18 42 60 3360 10 33 43 3370 26 43 69 3380 22 43 65 3390 3 28 31 3400 7 57 64 3410 19 46 65 3420 4 22 26 3430 9 54 63 3440 21 45 66 Total 577 862 1439

The pooled arcsine transformed values of the percent frequency of occurrence were correlated with the mean catch per tow for a given stratum (Figure E2). As abundance increases (the catch/tow is elevated), the species becomes more widely distributed in the core strata.

Table E8. Presence – absence of Atlantic croaker in trawl tows made during the southern leg of the NMFS – Woods Hole fall groundfish survey in the core strata as previously defined. Data pooled across all years (1972 – 2002).

100

Atlantic croakerNMFS Survey

Mean Number per Tow by Stratum

0 100 200 300 400 500 600 700 800

Arc

sin

(Per

cen

t O

ccu

rren

ce)

0

10

20

30

40

50

60

70

18 19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37 38

39 40

41

42 43

44

trend line

r = 0.7

Examination of the length frequency distribution of the Atlantic croaker by strata for the time series showed that smaller individuals (<5-cm TL) were consistently taken in a limited area within the core section. The average size of the Atlantic croaker in the individual strata was smallest in the same general area along the coast where the smallest size classes were most abundant (Tables E9 and E10). Their distribution would suggest that the origin of these small fishes off the eastern shore of Maryland and Virginia was Delaware Bay.

Figure E3 Relationship of the arcsine transformed frequency of occurrence and the mean catch per tow for Atlantic croaker taken during the southern leg of the NMFS-Woods Hole groundfish survey. Data derived from values pooled for each stratum over the time series. Numbers refer to stratum designations under old system, i.e., 38 = 3380.

101

Stratum TL 3180 3190 3200 3210 3220 3230 3240 3250 3260 3270 3280 3290 3300 3310 3320 3330 3340 3350 3360 3370 3380 3390 3400 3410 3420 3430 3440 1 5 210 2 6 2668 1870 70 1865 1782 144 3 1 16 1314 600 28 3039 1480 532 11 1 6 4 10 368 25 566 78 51 33 1 5 3 99 157 9 1 13 6 3 36 22 10 1 7 12 5 3 3 8 1 6 1 1 9 1 10 9 2 11 1 1 3 4 8 12 4 3 1 16 254 6 1 57 14 13 7 1 2 1 7 38 16 8 11 5 2243 21 835 75 14 15 1 21 2 21 19 166 18 4699 172 12 1155 379 29 15 74 6 44 112 19 47 10 4 1577 18 3120 323 15 1356 604 103 16 34 468 37 260 13 224 21 1 76 8 5 246 2 3079 105 5 2265 349 84 1091 1215 346 17 34 762 80 814 92 263 7 81 19 103 2405 12 4583 1051 35 1466 1056 334 621 2718 615 18 448 1 706 137 1002 120 167 17 70 84 955 7366 101 3549 2786 98 1630 1845 972 587 4141 1253 19 1654 2 2 382 98 504 155 124 45 45 337 256 62 2476 8267 599 2910 4072 220 1933 2615 1756 635 4692 1800 20 2381 1 2 134 105 373 177 229 180 249 690 1332 445 3449 5363 1785 3512 2678 813 2441 4285 2661 542 3685 1954 21 1864 4 1 27 52 173 108 668 787 329 875 1928 1361 1689 4747 2945 2930 1145 1187 2956 4063 2808 601 2643 1674 22 1833 2 1 43 60 153 125 748 1320 503 645 1810 2462 1541 3943 3317 1867 1754 1287 1329 2485 4149 1381 1325 951 23 594 2 1 2 4 151 143 201 892 1445 375 444 1674 2572 1000 3221 3811 715 2125 767 772 1783 3850 1895 837 873 24 138 4 1 1 50 328 183 275 618 1636 478 541 1321 2252 527 2483 2450 162 2285 406 524 1101 2266 1366 353 348 TL 3180 3190 3200 3210 3220 3230 3240 3250 3260 3270 3280 3290 3300 3310 3320 3330 3340 3350 3360 3370 3380 3390 3400 3410 3420 3430 3440

Table E9. Length frequency (total length in cm) distribution of Atlantic croaker by stratum for the southern leg of the NMFS-Woods Hole groundfish survey. Data pooled by stratum across all years. Note most small fishes <6 cm TLs were taken in strata 3270, 3280, 3300 and 3320. These are the two shallowest strata south of Delaware Bay off the eastern shore of Virginia and Maryland.

102

25 140 5 2 9 135 641 192 309 465 1352 865 649 1311 1469 254 1014 2286 427 1584 307 221 706 909 817 393 160 26 74 46 3 12 264 17 982 248 431 332 1110 827 421 1078 1175 141 662 1384 42 1090 125 143 565 312 479 316 73 27 47 211 7 32 256 53 927 419 448 255 736 852 152 1205 877 62 396 958 18 682 84 65 409 113 355 100 79 28 80 300 15 53 189 102 650 478 466 123 525 863 152 728 454 17 287 510 19 402 53 101 309 35 76 205 84 29 54 472 21 85 230 78 534 545 231 85 341 700 45 417 253 6 178 337 36 162 39 15 210 19 90 31 30 64 523 26 89 176 90 387 573 214 59 320 668 54 237 153 4 109 289 7 41 27 69 185 23 51 47 47 31 18 599 15 98 112 108 277 309 194 32 116 248 11 110 106 5 30 153 1 63 12 63 52 17 51 55 33 32 83 469 19 93 134 102 222 396 66 21 78 274 21 40 65 5 16 89 5 19 4 66 9 1 9 33 111 415 11 51 97 97 97 305 55 14 39 209 3 24 89 6 25 48 8 17 10 45 6 13 28 34 140 134 2 26 45 64 44 219 29 4 45 121 3 38 65 1 18 69 1 4 4 37 12 14 1 35 36 98 2 13 28 37 39 191 30 2 76 1 1 42 11 31 18 9 21 5 3 4 36 1 77 2 6 20 41 17 34 16 1 5 29 1 5 43 7 13 3 6 1 13 2 37 4 1 6 27 7 43 9 44 1 19 1 8 10 1 2 5 4 38 35 9 25 3 24 4 26 2 21 2 23 39 3 18 1 25 3 13 4 17 6 4 40 1 3 4 4 2 20 32 1 2 1 1 41 2 1 1 9 1 2 1 5 3 1 42 1 13 1 2 2 7 43 2 8 3 3 44 1 2 1 6 45 1 1 1 46 1 1 47 3 15 48 2

103

Stratum Number Mean TL cm std error 95% Confidence Interval lower upper

3180 9865 21.4 0.033 21.4 21.5 3190 3381 30.7 0.041 30.6 30.8 3200 136 29.7 0.298 29.1 30.3 3210 3194 19.9 0.092 19.7 20.1 3220 1762 28.8 0.071 28.7 28.9 3230 882 31.7 0.118 31.5 31.8 3240 5929 26.5 0.052 26.4 26.6 3250 7660 24.7 0.071 24.5 24.8 3260 3785 25.8 0.069 25.6 25.9 3270 9971 13.5 0.103 13.3 13.7 3280 12852 19.9 0.084 19.7 20 3290 7915 26.5 0.052 26.4 26.6 3300 11057 12.4 0.096 12.2 12.6 3310 17007 19.6 0.068 19.4 19.7 3320 14075 24.1 0.026 24.1 24.2 3330 13012 19.6 0.04 19.6 19.7 3340 40806 20.5 0.013 20.4 20.5 3350 21248 23.4 0.02 23.4 23.5 3360 25670 18.7 0.016 18.7 18.7 3370 22110 21.5 0.021 21.4 21.5 3380 5526 22.2 0.033 22.1 22.2 3390 26333 17.5 0.022 17.5 17.6 3400 22731 21.1 0.021 21 21.1 3410 20370 21.7 0.015 21.7 21.7 3420 13963 19.8 0.037 19.7 19.9 3430 23925 19.4 0.018 19.4 19.4 3440 10495 20.3 0.026 20.3 20.4

total 355660 20.6 0.009 20.6 20.6 Age determination and age composition In sciaenids (drums and croakers), the major problem in the determination of ages from otoliths is the definition of the first annulus. This becomes more difficult when the species either has a protracted spawning season over its geographical range or spawns in the late summer-early fall. For example, the red drum (Sciaenops ocelatus) is a late summer spawner in South Carolina waters. The young enter the tidal creeks for a two to three month period and move to deeper waters in the winter at a size of 3 to 5 cm TL. When the waters warm in the spring, the young move back into the shallow tidal creeks where they feed on a variety of crustaceans such as grass shrimp. They reach a size of 15

Table E10. Mean total length (cm) of Atlantic croaker taken in the southern leg of the fall groundfish survey conducted by NMFS-Woods Hole. Data pooled over the time series by stratum within the core area.

104

to 20 cm TL by June when they leave this habitat and form schools of like sized individuals in various shallow water areas of the estuary. By their second winter, they reach a size of 30 to 40 cm TL. During the cold weather period, feeding is reduced and the growth is essentially zero. The following spring (April – early May) these individuals deposit the first well defined mark on the edge of their sagittae. These fishes are from 20 to 21 months of age at the time of formation of the first well-defined annulus. Careful examination of the otolith’s core in transverse sections frequently shows a series of ill-defined, hazy concentric rings near the core. This probably is a combination of the settlement mark as well as passage through the first winter. The distance from the center of the core to the position of these rings is highly variable, and in some individuals, the area (the diffuse rings) is entirely absent. Since the annual rings on the otoliths of red drum are very distinct, we ignore the ill-defined “marks” near the core and designate the first well defined ring as the first annulus. Red drum, as previously mentioned, are ~20 – 21 months old at this time. The situation is even more complex in the Atlantic croaker. I followed the embedding and sectioning procedures outlined for this species in a draft document dealing with age determination of fishes in the Gulf of Mexico produced by the Gulf States Marine Fisheries Commission (Table 9). Ootliths (sagittae) were removed, stored dry and later sectioned (~0.5-mm) with a low speed saw after embedding in expoxide resin. Transverse sections were examined under a binocular microscope and the ages determined. . Atlantic croaker Micropogonias undulatus Highlights _ Otoliths relatively easy to locate and extract. _ Multiple sectioning techniques successful. _ Rings easily discernable. _ First distinct opaque ring forms at approximately 1.5 years of age. _ Generally less than ten rings. Otolith description The sagittae in Atlantic croaker are very thick and shield shaped, often with a shelf or flange on the outer surface or on the dorsal margin (Figure 5.75 ). The ostium of the sulcus is large, pear-shaped, and its expanded part does not reach the anterior margin. The ‘J’ shaped cauda of the sulcus acousticus is sharply bent, and its dorsal edge extends further into the ostium than its ventral edge. Otolith processing Due to the robust nature of the otoliths in this species, multiple techniques are acceptable and usually reflect available equipment. Generally, Atlantic croaker sections are processed at approximately 0.5 mm. The following techniques have been used

Table E11. Methodology for the determination of the ages of Atlantic croaker used in the present report. Extracted from GSMFC manual dealing with age determination of fishes in the Gulf of Mexico. Description directly copied from manual.

105

successfully throughout the Gulf. Low Speed Wafering Saw Techniques Embedded Whole Otoliths (Section 3.4.2.1) LDWF, GCRL, MDMR, FMRI 1. Embed the otolith with the long axis parallel to the long axis of the mold. 2. Locate core and position block in chuck. 3. Adjust arm weight and speed. Make successive 0.5 mm cuts to obtain the core region. 4. Mount the core sections. Mounted whole otoliths (Section 3.4.2.2)FMRI 1. Mount whole otolith to slide, concave side down with the long axis parallel to the long side of the slide using thermoplastic. 2. Locate core and position slide in chuck. 3. Adjust arm weight (50-75 g) and speed (8- 10). Make successive 0.5 mm cuts to obtain the core region. 4. Mount the core sections. High speed wafering saw techniques Embedded Whole Otoliths (Section 3.4.2.1) TPWD 1. Embed the whole otolith with the long axis parallel to the long axis of the mold. 2. Locate core and position block in chuck. 3. Adjust load (1,000 g) and speed (3,000 rpm). Make successive 0.5 mm cuts to obtain the core region. 4. Mount the core sections. Thin section machine Free-Hand whole otolith sectioning (Section 3.4.3) LSU, AMRD 1. Firmly grasping both ends of the otolith, make initial cut adjacent to the core. 2. Hand grind additional material until core is visible. 3. Mount otolith half with core on labeled slide. 4. Place slide in chuck and section off remaining material. 5. Place slide into precision grinder arm and adjust caliper to 0.5 mm. Age Determination Transverse otolith sections of Atlantic croaker show very clear, easily identified marks that can be used for aging. Typical sections have an opaque core surrounded by a blurred opaque band, composed of fine opaque and translucent zones. This band represents the first annulus. Because of Atlantic croaker’s spawning season, the width of the first annulus varies among individuals. Spawning typically occurs from November through January while annuli deposition occurs from December through May. Late-spawned fish have a very narrow band that is almost continuous with the core; early spawned fish have a wide, well-defined band clearly separated from the core. Because of this variation in width and proximity to the core, the first annulus is sometimes difficult to identify. Figure 5.80 Otolith section of an

106

age-8 Atlantic croaker. Black arrows indicate annuli. Note first annulus appears as a blur or smudge. Subsequent annuli are represented by easily identified, narrow, opaque bands that alternate with wider translucent bands outside the proximal margin of the first annulus. For regional stock assessment purposes, three minimal parameters are recorded: number of rings, presence or absence of an opaque ring at the margin, and month of capture. Based on these three parameters, cohort and biological ages can be determined. Other aging methods Whole otoliths have not been used successfully in the Gulf region. The usefulness of break and burn techniques for Atlantic croaker has not been determined, however; this species may be a good candidate for the technique. Atlantic croaker scales have not been demonstrated to be useful in the Gulf yet (Figure E4) Birthdate assignment timeline for Atlantic croaker. Age or year group based on biological birthdate (January 1), number of rings, and January 1 to December 31 year. In our estimates of the ages of Atlantic croaker, we used the methodology as described above for the embedding and sectioning (low speed). There is still a problem with the determination (using the GSMFC definition) of the first annulus. Near the core of many individuals (they define them as early spawned fish), a faint ring can be seen that is of variable location and size as well as clarity. From the description in the aging manual “Spawning typically occurs from November through January while annuli deposition occurs from December through May (Figure E5). Late-spawned fish have a very narrow band that is almost continuous with the core; early spawned fish have a wide, well-defined band clearly separated from the core. Because of this variation in width and proximity to the core, the first annulus is sometimes difficult to identify. Otolith section of an age-8 Atlantic croaker. Black arrows indicate annuli. Note first annulus appears as a blur or smudge. Subsequent annuli are represented by easily identified, narrow, opaque bands that alternate with wider translucent bands outside the proximal margin of the first annulus.” We have not included this area (blur or smudge) in the counts of the annuli because it is not present in all fish from either the Middle or South Atlantic. Photomicrographs of Atlantic croaker sectioned otoliths used in our analysis are in Figures E4 and E5. Spawning season and origin of problems in age determination The confusion in the designation of the first annulus results from the extended spawning season for this species along the eastern United States. In the Chesapeake Bay region, the spawning dates for larval and juvenile Atlantic croakers were from early July to early February with an estimated 82% of spawning occurring from August to October (Nixon and Jones 1997). In the Cape Lookout area of North Carolina, Warlen and Burke (1990) found Atlantic croaker move from coastal to estuarine waters over a six-month period from November through April. They defined two pulses of larval ingress. The first

107

occurred in the fall (November-December) and the second was in mid-February to mid-April. Plankton samples taken during the six weeks between these two peaks had a much lower abundance of Atlantic croaker (Warlen and Burke 1990). In the South Atlantic Bight, data are available for the Cape Fear River estuary (Weinstein 1979), South Carolina (McGovern and Wenner 1990; Wenner, unpublished data). In the Cape Fear River, Weinstein sampled the shallow tidal creeks that meander through the vegetated marsh systems up a salinity gradient. Atlantic croaker ingress to the shallow creeks peaked in November, but generally the catches were low. He indicated that the species preferred the deeper waters of the estuarine system over a muddy bottom rather than the shallow tidal creek system (Weinstein 1979). In a subsequent study in that same area, Weinstein et al. (1980) determined that ingressing Atlantic croaker were most concentrated in the near bottom waters of the Cape Fear River. These fishes ranged in size from 7 to 30 mm SL with a modal size of 11 mm. The interesting fact about these collections was that they were made over a three two-day periods in 1977 (March 14-15, April 5-6, April 11-12) and they contained a number of larval and early juvenile Atlantic croaker that were not seen in the samples from the shallow marsh habitat of the same river system. They failed to give any indication as the period of peak ingress (summer-fall; winter) Further south in the North Inlet estuarine system, Bozeman and Dean (1980) sampled an intertidal creek with a blocknet thereby capturing all fishes that had moved into the area on the high tide. These collections were made from October 21, 1974 to February 22, 1975. Atlantic croaker were present in all collections except the final sample taken on May 25, 1975. Although no lengths were presented for the fishes, the number and weight of each species was taken (Table E12). As can be seen from the average weights in the right hand column of Table E12, these were small fishes suggesting that recruitment of Atlantic croaker into this system occurs over an extended period from October through February. In this same system with another gear type (epibenthic sled), Allen and Barker (1990) found increases in abundance in late fall and early spring with more Atlantic croaker in the latter. The average size of these fishes was from 9 to ~15 mm SL. In the Charleston Harbor estuarine system, plankton samples caught Atlantic croaker from October through May. Greatest catches were in January, and the late winter-early spring tows were far higher than those in the fall (Wenner, unpublished data). These were small fishes with most being less than 15 mm TL.

108

Date Number Weight Average weight 3 October 2,532 31.4 0.012 2 November 166 2.5 0.015 14 December 92 1.9 0.021 4 January 4 0.1 0.025 7 January 3,425 92.4 0.027 16 January 10 0.4 0.040 30 January 2,187 67.7 0.031 22 February 323 12.2 0.038 8 March 410 16.8 0.041 15 March 124 4.8 0.039 17 March 86 3.2 0.037 20 April 14 0.7 0.050 25 May 0 0

In yet another plankton study, McGovern and Wenner (1990) indicated that Atlantic croaker showed a “protracted period of recruitment to the creek habitat, with larvae and small juveniles found in collections from September through May”. This extended period of larval ingress that varies latitudinally is the source of the difficulty in the determination of the first annulus. The first well defined ring outside of the core area (with its concentric, multiple rings and ‘smudges’) is designated as the first annulus. This is deposited in the spring (April-May in South Carolina). Hence, the age at the first well defined ring is from 13 to 18 months. By ignoring the noise near the core, much of the confusion is eliminated as the remainder of the marks on the structure are readily interpretable. The purpose of this lengthy description of larval ingress into the nursery habitat is to describe the origin of the problems with the designation of the first annulus. By ignoring the noise around the core and using only well defined rings as an interpretation of the age, the ages can be determined more accurately. Since there is recruitment to east coast estuaries over an extended period with peaks dependent upon latitude, by ignoring the core and designating January 1 (the approximate mid-point of the spawning season) as the birthdate, age determination for this species is standardized. Fishes sampled during the NMFS-Woods Hole fall groundfish survey (length based subsample of fishes in each tow) were aged and placed in 1-cm size intervals. An age-length key was constructed for the survey. Note that the available ages were survey specific but not cruise specific, i.e., ages were not available for every cruise. We have unprocessed otoliths from the last four years. The methodology used assumes that there are not yearly changes in the lengths at age.

Table E12. Total number and weight and mean weight of Atlantic croaker in samples from North Inlet, SC as reported in Bozeman and Dean

109

In the photomicrograph figures, the characteristics of the rings near the core are visible. In Figure E4, there is a translucent “ring” near the center. This is surrounded by a series of rings that are of variable clarity. In our age determinations, the “ring” was not included in the counts of the annuli. This fish was designated as being age 4 +. Figure 5 shows what we call an age 0 fish taken during the fall groundfish survey along the Middle Atlantic Coast during September. It would have deposited its first well defined ring the following spring.

The construction of the age-length key for the NMFS fall survey was based on the ages of fishes captured with the survey gear during regular cruises in the mid to late 1990’s. No data are available for the prior years. Otoliths are available, but as yet have not been read, from cruises made from 1999 through 2003. Analysis is planned on these otoliths in the near future as well as participating in the fall surveys of 2004 and 2005. The

Figure E4. Photomicrograph of a transverse section of a sagittal otolith from an Atlantic croaker collected during the southern leg of the fall groundfish survey of the Middle Atlantic coast. We have interpreted the age of this fish as 4+. The ring close to the core and the smudge near it were not included in the count in contrast to methodology in GSMFC handbook on age determination.

110

following results are based on keys produced from a few years merged together and applied to the time series of catches and lengths from 1972 through 2002. The resulting estimates of the age composition of the catch during the NMFS fall survey are in Table E12. The oldest fishes (age 9 – 10) were present only in the mid-1970’s and recent years. The periods of expanded age distributions coincided with peaks in the commercial landing along the east coast of the United States. In addition, the NMFS-Woods Hole fall survey estimates of abundance as indicated by the stratified mean catch per tow of the core strata were high during the same periods. Given the habit of young-of-year Atlantic croaker (resulting from late summer spawning events) over-wintering

in some estuaries such as Chesapeake Bay, the abundance estimates of the age 0 fish do not show the same trends, i.e., fails to track with the commercial landings, because of the unavailability to the survey gear. Perhaps the indices of abundance derived from other surveys inside the various estuaries during the fall and winter could provide a much better

Figure E5. Photomicrographs of a cross section of a sagittal otolith from and Atlantic croaker collected during the fall groundfish survey along the Mid-Atlantic coast. The multiple rings near the center were not defined as the first annulus. The fish was designated as age 0. Magnification greater than Figure 4.

111

estimate of yearclass strength. The values could be lagged for a year and then compared to the abundance of say, age 1 fish in the NMFS survey. The contributions of the various age groups to the stratified mean catch per tow for a species year’s survey are in Table 11. Higher values of the older age groups (age 4+) are in the more recent year and in the 1970’s. Future Work The SCDNR plans to complete the age determinations from the NMFS-fall survey off the Mid-Atlantic coast as well as complete the gathering and analysis of the SEAMAP South Atlantic data. The project has taken fish for age analysis for the past five or more years, so at least for the most recent period, there are cruise specific age-length keys.

112

cruise age 0 age 1 age 2 age 3 age 4 age 5 age 6 age 7 age 8 age 9 age 10 total

197208 276 9 3 288197308 661 1333 527 212 53 22 8 2 2818197411 396 2766 2026 1446 501 226 90 50 5 1 3 7510197512 29277 17258 3683 1779 383 218 76 53 9 1 3 52740197609 1860 9856 3683 2215 596 310 114 63 12 1 1 18711197712 214 2048 1302 1192 350 211 114 59 26 1 4 5521197806 471 4372 2193 1403 484 274 130 74 34 3 5 9443197910 308 287 66 46 14 8 5 2 2 738198007 557 2423 580 206 43 18 9 5 3 3844198106 227 717 230 105 28 10 6 1 1324198206 6 110 81 62 15 7 4 1 286198306 5382 5592 687 132 9 2 11804198405 1310 7366 2373 782 190 75 25 3 12127198508 9118 3722 4291 839 244 57 22 9 2 1275198606 737 2772 1027 375 89 61 11 7 5079198705 245 2321 869 211 65 19 3 3733198803 66 701 279 77 22 1 1 1147198904 1779 1916 412 104 18 5 1 4235199004 2367 1299 205 42 3 3916199105 6144 7397 968 204 28 4 6 1 14752199206 3463 5269 885 275 24 5 0 0 9921199306 8600 1942 478 135 26 5 1 11187199406 4626 10291 3052 1098 239 111 28 12 19457199507 5072 3108 749 232 50 21 7 9239199604 1650 2481 1801 1697 512 310 128 70 3 1 1 8654199706 4329 2176 704 577 152 100 40 24 1 8103199804 5388 7019 2154 986 257 123 60 25 5 1 2 16018199908 4958 14831 5959 3137 890 498 183 109 18 5 4 30592200005 4109 7655 3285 1849 512 287 111 56 9 1 1 17875200109 951 2128 1057 2175 841 547 261 184 43 3 14 8204200209 20780 17952 3544 2874 1000 643 302 193 44 9 18 46825

Table E13. Age composition of the southern leg of the fall groundfish survey conducted by NMFS-Woods Hole. Ages estimated by the application of an age-length key (determined by sections of sagittae) to the expanded length frequency distributions of each cruise. Reading procedures of the sections followed those of the Gulf.

113

cruise Total Number sty % - 0 % - 1 % - 2 % - 3 % - 4 % - 4+

197208 288 2.33 95.83 3.13 1.04 0.00 0.00 0.00197308 2818 38.07 23.46 47.30 18.70 7.52 1.88 1.14197411 7510 143.20 5.27 36.83 26.98 19.25 6.67 4.99197512 52740 638.21 55.51 32.72 6.98 3.37 0.73 0.68197609 18711 397.61 9.94 52.67 19.68 11.84 3.19 2.68197712 5521 119.35 3.88 37.09 23.58 21.59 6.34 7.52197806 9443 161.72 4.99 46.30 23.22 14.86 5.13 5.51197910 738 15.64 41.73 38.89 8.94 6.23 1.90 2.30198007 3844 88.53 14.49 63.03 15.09 5.36 1.12 0.91198106 1324 31.77 17.15 54.15 17.37 7.93 2.11 1.28198206 286 9.11 2.10 38.46 28.32 21.68 5.24 4.20198306 11804 231.94 45.59 47.37 5.82 1.12 0.08 0.02198405 12127 267.61 10.81 60.76 19.57 6.45 1.57 0.85198508 1275 213.97 49.81 20.33 23.44 4.58 1.33 0.49198606 5079 127.11 14.51 54.58 20.22 7.38 1.75 1.56198705 3733 111.96 6.56 62.18 23.28 5.65 1.74 0.59198803 1147 31.65 5.75 61.12 24.32 6.71 1.92 0.17198904 4235 99.64 42.01 45.24 9.73 2.46 0.43 0.14199004 3916 79.82 60.44 33.17 5.23 1.07 0.08 0.00199105 14752 260.53 41.65 50.14 6.56 1.38 0.19 0.07199206 9921 216.19 34.91 53.11 8.92 2.77 0.24 0.05199306 11187 140.88 76.87 17.36 4.27 1.21 0.23 0.05199406 19457 478.57 23.78 52.89 15.69 5.64 1.23 0.78199507 9239 189.36 54.90 33.64 8.11 2.51 0.54 0.30199604 8654 203.99 19.07 28.67 20.81 19.61 5.92 5.93199706 8103 159.14 53.42 26.85 8.69 7.12 1.88 2.04199804 16018 344.79 33.63 43.81 13.45 6.15 1.60 1.35199908 30592 734.45 16.21 48.48 19.48 10.25 2.91 2.67200005 17875 387.65 22.99 42.83 18.38 10.34 2.86 2.60200109 8204 177.64 11.59 25.94 12.88 26.51 10.25 12.82200209 46825 939.82 44.57 37.78 7.10 5.96 2.04 2.55

Table E14 . Percent of the number of Atlantic croaker caught during the southern leg of the

fall groundfish survey in each age group; sty is the stratified mean catch per tow in

numbers, total number is the total number of Atlantic croaker caught during the cruise; % - 0 = the percent of the total number of Atlantic croaker at age 0 and so on; % - 4+ is the percent of the total that is greater or equal to age 5.

114

Literature cited ALLEN,D.M. AND D.L. BARKER. 1990.

Interannual variations in larval fish recruitment to estuarine epibenthic habitats. Mar. Ecol. Prog. Ser. 63: 113-125.

AZAROVITZ, T.R. 1994.

Northeast Fisheries Science Center Bottom Trawl Surveys. In Proceedings of the workshop on the collection and use of trawl survey data for fisheries management, p. 4-7. ASMFC Special ReptNo.35

BOZEMAN, E.L.,Jr. AND JM. DEAN. 1980.

The abundance of estuarine larval and juvenile fish in a South Carolina intertidal creek. Estuaries 3: 89-97.

GROSSLEIN, M.G. 1969.

Groundfish survey program of BCF Woods Hole. Commercial Fisheries Rev. (8-9) 22-35.

KREBS, C.J..1989. Ecological Methodology. Harper Collins, N.Y., N.Y., 654 p. McGOVERN, J.C. AND C.A. WENNER. 1990.

Seasonal recruitment of larval and juvenile fishes into impounded and non-impounded marshes. Wetlands 10: 203-221.

NIXON, s.w. AND C.M. JONES. 1997.

Age and growth of larval and juvenile Atlantic croaker, Micropogonias undulatus, from the Middle Atlantic Bight and estuarine waters of Virginia. Fish. Bull. 95: 773-784.

WARLEN, S.M. 1982.

Age and growth of larvae and spawning time of Atlantic croaker in North Carolina. Proc. Annual Conf. South eastern Assoc. Fish & Wildl Agencies. 34: 204-214.

WARLEN,S.M. AND J.S. BURKE. 1990.

Immigration of larvae of fall/winter spawning marine fishes into a North Carolina estuary. Estuaries 13: 453-461.

WEINSTEIN, M.P. 1979.

Shallow marsh habitats as primary nurseries for fish and shellfish, Cape Fear River, North Carolina. Fish. Bull. 78: 419-436.

115

WEINSTEIN, M.P., S.L. WEISS, R.G. HODSON, AND L.R. GERRY. 1980. Retention of three species of postlarval fishes in an intensely flushed tidal estuary, Cape Fear River, North Carolina. Fish. Bull. 78:419-436.

WENNER, C.A., C.A. BARANS, B.W. STENDER, AND F.H. BERRY 1979.

Results of MARMAP otter trawl investigations in the South Atlantic Bight. I. Fall, 1973. Tech. Rept. 33, S.C. Mar. Resour. Cent., Charleston, SC 29412, 78 p.

116

Appendix F: An Evaluation of Weighting the Likelihood terms in an Age Structured Production Model for Atlantic croaker

Data weighting has been identified as a crucial problem in modern stock assessments (NRC 1998). One of the recommendations of the review panel for Atlantic croaker was to examine the weighting of the likelihood components in greater detail. Weighting of the likelihood components has the potential to have a significant influence of the model outcome. In this report we examine possible weighting options for the Atlantic croaker model and attempt to objectively determine a suitable suite of likelihood weightings for use in the model implementation. In the original version of the age structured production model, we gave the fleets, recruitment deviations and MRFSS index a weight of λ =1, and all fishery independent indices a weight of λ=2. In this iteration of the model we explore alternate weighting schemes in more detail. The likelihood components fall into three groups, the fleet, index, and the recruitment deviation components. The fleet and index likelihood terms are based on the difference between the observed data and the predicted estimates. The recruitment deviation likelihood is based on differences from a mean of 0 (i.e. no deviation from the stock-recruit relationship). As such, weightings were treated in two groupings: 1) weights for the fleets and indices; and 2) weights for the recruitment deviations. Weighting profiles for the fleets and indices were examined while keeping the weight on the recruitment deviations constant at λ=1. All of the likelihood terms were estimated assuming a lognormal distribution. To determine suitable weights for the fleets and indices we used a set of criteria to rank the relative importance/reliability of each fleet to each other and each index relative to the other indices. For the fleets, we ranked each fleet on two criteria:

1) The number of years for which data were available. 2) An estimated proportion of total landings that was determined using the available data.

For the indices we ranked each of the indices on three criteria:

1) Number of years sampled. 2) Sampling design (fishery independent/dependent)/geographical coverage. 3) Seasonal coverage of sampling within a year.

We chose three base weights that were used in a combinatorial design, to evaluate 16 weighting combinations. The four base weights were λ =3.37 (~c.v. = 0.4), λ =4.33 (~c.v. =0.35), λ =5.8 (~c.v. =0.3) and λ = 8.25 (~c.v. =0.25). In addition, we also ran the model where all weights were set to λ =1, one where the indices were set to λ =2, and the weighting method used in the previous assessment (fishery independent indices λ =2, fleet and fishery dependent indices λ=1). Table F1 summarizes the relative rankings and the weighting schemes evaluated.

117

To evaluate the influence of weighting, the performance statistics used were the standard deviations of the residuals for each of the indices and fleets, the total un-weighted likelihood of each of these terms, the recruitment deviation likelihood and the total un-weighted likelihood (Tables F3 and F4). We did not consider using the parameter estimates of the model (e.g. fishing mortality rates) as good indicators as it would be difficult to evaluate them objectively. In our ranking system, the NEFSC, SEAMAP, and VIMS index received similar relative weightings and the MRFSS index ranked the lowest. In general, increasing the weight on a likelihood component resulted in a reduction in the standard deviation of the residuals. However, with the exception of the SEAMAP index the observed reductions were relatively small (Figures F1 to F4). For the fleets, the commercial and recreational fishery received the highest relative rankings, with the scrap/discard and shrimp bycatch having low relative rankings. For the fleets, increasing λ resulted in a more marked reduction in the standard deviation of the residuals (Figures F5 to F8), indicating a better fit to those terms could be obtained. When examining the effects of weighting options on the total likelihood, it was readily apparent that increasing the weighting on all four indices had a similar effect (Figures F9-F12). The similar trends are to be expected as the weighting of the indices are tied together through their relative rankings. In general, increasing the weightings, reduces the likelihood terms for the indices (these are the un-weighted values), but is compensated by an increase in the recruitment deviation and fleet likelihood terms. For the NEFSC and SEAMAP indices weightings > 4 appear to produce relatively small reductions in the total likelihood of the indices (Figures F9 and F11). Trends in the relationship between the un-weighted likelihood and weightings among the different fleets were similar (they too were linked by their relative ranking). For the fleets, increasing the weighting on the likelihood results lowering the fleet likelihood, with the recruitment deviation and index likelihood terms remaining relatively flat across most weighting combinations. It should be noted that for each of the fleets, the model estimates 30 parameters (a fully selected F for each year), whereas for the indices the model estimates one parameter per index (a catchability coefficient). As such, one would expect the fleets to provide a better fit, irrespective of weighting. This is reflected in the standard deviation of the residuals (Table F3). Comparing the total un-weighted likelihood terms for the 19 model runs suggests that for most runs the total likelihood’s were similar (Figure F17). In fact, the base run where all components were weighted λ=1, and the original weighting used in the previous version, produced the lowest un-weighted total likelihood. Model runs 15, 10, 5, 14, 9, 13, ‘Original’ and ‘index x 2’ had likelihood terms that were at most 15 % greater than the model run with the lowest total likelihood (the base run). With the exception of the ‘Original’ and ‘index x 2’ runs, all other runs had weighting schemes that favored the fleets over indices.

118

There appears little evidence to suggest that any of our weighting schemes make a meaningful reduction in the total un-weighted likelihood. However, increasing the weighting on an index or fleet will reduce the standard deviation of that component and its respective likelihood component. For the indices, it appears that forcing the model to fit the indices better results in compensation through adjusting the recruitment residuals to account for the greater confidence in those indices. It is possible that the upper bound for our weighting choices were too low. We re-did the analysis using higher base weights (λ= 10, 15 and 20). The results of that analysis were similar to the weighting scheme described in the text. One of the disadvantages of the relative ranking of the fleets and indices is that it ties them together. A combinatorial type design, where all possible combinations of a set of weights are examined, would be an appropriate method to explore for the future. As an alternative, we ran a simulation of 5,000 runs where each weighting term was randomly and independently assigned a weight between 0 to 20 using a uniform random number generator. The results of the 5,000 simulations were enlightening. None of the 5,000 simulations had a total un-weighted likelihood less than the “base” and “original” models. There also appears to be a strong indication of a flat response surface, 60% of runs had very similar total likelihood’s (Figure F18). Increasing the weighting terms reduced the standard deviation of the residuals for the NEFSC, SEAMAP and VIMS index and had less influence on he MRFSS index and fleets (Figures F19-F26). In our examination of weighting terms, we have treated the recruitment deviations separate to the other likelihood terms. It is apparent that increasing the weight on an index is compensated by increased recruitment deviations. Preliminary runs suggest that the recruitment deviations show a correlation with those indices that have a large age 0 component. Furthermore, it is difficult to objectively determine an appropriate weighting. Maunder and Deriso (2003) through a series of simulations noted that for New Zealand snapper that a standard deviation of 0.6 was appropriate. In terms of λ, a standard deviation of 0.6 ~ 1.39, which is relatively close to our weighting of 1. Maunder and Deriso (2003) note that Beddington and Cooke (1983) found that the standard deviation of recruitment residuals for many species of fish was around 0.6. However, none of species noted in Beddington and Cooke (1983) had similar life history characteristics to Atlantic croaker. We examined the effects of increasing the weighting of the recruitment residuals on two fleet-index weighting combinations. For the fleet-index weight combinations, we chose the base run (all λ =1) and the ‘original’ weight combination (fishery independent indices had a λ =2 and all other terms λ =1). For the recruitment deviation weights we chose six weights. The base case was a λ =1, Other choices were λ =1.39 ~ sd=0.6 and similar to that of Maunder and Deriso (2003), λ = 3.37, 4.33, 5.8 and 8.25, values that were used for evaluating the fleet and index weighting options. Tables F4 and F5 summarize the affects of increasing the recruitment deviation weights. To summarize, increasing the weighting on the recruitment deviations, results in increasing the index and fleet likelihood’s (poorer fit) and thus the total likelihood. Increasing the recruitment deviation weights constrains the model more closely to the stock-recruitment relationship. Weighting values similar to that suggested by the literature produced similar results to weightings used in the ‘original’ run. Choosing an appropriate weighting for the

119

recruitment residuals should be based on our interpretation on how closely the model should follow the stock-recruitment relationship. In this version of the model, steepness is a fixed value. Preliminary runs that estimated steepness using the revised data produced steepness estimates of 1, much higher than the base case of 0.76 used. Conclusions Our examination of possible weighting options revealed a relatively flat response surface for the likelihood terms. It was evident that none of the weightings considered produced a fit better than the base model. Simulations indicated that increasing an individual weighting component > 5 produced relatively little reduction in the standard deviation of the residuals. We were not able to objectively determine an appropriate weighting scheme. However, subjectively, we believe that, the fishery independent indices should be given a higher weight that the fleets. Our original weighting scheme appears to be a reasonable choice for the data. Literature Cited Beddington, J.R. and Cooke , J.G. 1983. The potential yield of fish stocks. FAO Fish. Tech. Paper, 242: 1-47. Maunder, M.N. and R. B. Deriso. 2003. Estimation of recruitment in catch-at-age-models. Can. J. Fish. Aquat. Sci. 60: 1204-1216. National Research Council. 1999. Improving fish stock assessments. National Academy Press.

120

Table F1. Relative rankings and weighting schemes explored for fleet and index likelihood terms. LANDINGS INDICES Rec dev

Commercial Recreational Scrap/discardShrimp bycatch NEFSC MRFSS SEAMAP VIMS

Relative. Rank 1.00 0.59 0.19 0.13 1.00 0.37 0.93 0.75Base 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1Index x 2 1.00 1.00 1.00 1.00 2.00 2.00 2.00 2.00 1Original 1.00 1.00 1.00 1.00 2.00 1.00 2.00 2.00 1run 1 3.37 1.98 0.65 0.45 3.37 1.24 3.14 2.53 1run 2 3.37 1.98 0.65 0.45 4.33 1.59 4.04 3.25 1run 3 3.37 1.98 0.65 0.45 5.80 2.13 5.41 4.35 1run 4 3.37 1.98 0.65 0.45 8.25 3.02 7.70 6.19 1

run 5 4.33 2.54 0.83 0.58 3.37 1.24 3.14 2.53 1run 6 4.33 2.54 0.83 0.58 4.33 1.59 4.04 3.25 1run 7 4.33 2.54 0.83 0.58 5.80 2.13 5.41 4.35 1run 8 4.33 2.54 0.83 0.58 8.25 3.02 7.70 6.19 1run 9 5.80 3.40 1.11 0.77 3.37 1.24 3.14 2.53 1run 10 5.80 3.40 1.11 0.77 4.33 1.59 4.04 3.25 1run 11 5.80 3.40 1.11 0.77 5.80 2.13 5.41 4.35 1run 12 5.80 3.40 1.11 0.77 8.25 3.02 7.70 6.19 1run 13 8.25 4.84 1.58 1.10 3.37 1.24 3.14 2.53 1run 14 8.25 4.84 1.58 1.10 4.33 1.59 4.04 3.25 1run 15 8.25 4.84 1.58 1.10 5.80 2.13 5.41 4.35 1run 16 8.25 4.84 1.58 1.10 8.25 3.02 7.70 6.19 1

121

Table F2 Summary of Likelihood terms for model weighting schemes evaluated. Filled cells indicate lowest likelihood for term.

Index likelihood

Fleet likelihood

Total likelihood

Run NEFSC MRFSS SEAMAP VIMS COMM RECScrap/dis

c Shrimp Fleet + Index Rec. dev TotalBase 10.69 4.25 2.93 12.14 0.19 0.02 0.45 1.00 31.67 9.60 41.27Index x 2 9.96 4.08 2.64 9.90 0.27 0.04 0.87 1.14 28.91 13.73 42.64Original 9.95 4.75 2.55 9.68 0.27 0.04 1.09 1.04 29.37 13.31 42.68run 1 7.56 4.94 2.42 10.32 0.05 0.02 3.37 6.32 35.01 13.92 48.94run 2 7.21 4.94 2.39 9.57 0.07 0.03 3.81 9.85 37.86 15.59 53.45run 3 6.90 4.91 2.35 8.74 0.09 0.04 4.48 14.76 42.27 17.73 60.00run 4 6.14 5.37 2.35 7.65 0.17 0.07 16.90 16.37 55.02 20.41 75.43run 5 7.92 4.85 2.42 10.56 0.03 0.01 2.42 3.73 31.94 14.35 46.29run 6 7.49 4.87 2.39 9.94 0.04 0.02 2.96 6.34 34.06 15.79 49.85run 7 7.08 4.89 2.36 9.11 0.06 0.03 3.42 11.15 38.10 17.82 55.92run 8 6.74 4.87 2.32 8.19 0.10 0.04 4.36 17.84 44.45 20.50 64.95run 9 8.32 4.70 2.41 10.68 0.02 0.01 1.41 2.09 29.64 15.03 44.67run 10 7.98 4.70 2.38 10.23 0.02 0.01 1.88 3.19 30.39 16.32 46.72run 11 7.43 4.79 2.36 9.60 0.04 0.02 2.54 6.67 33.44 17.90 51.33run 12 6.99 4.84 2.34 8.64 0.06 0.03 3.15 12.83 38.88 20.30 59.18run 13 8.61 4.60 2.40 10.77 0.01 0.00 0.70 1.22 28.30 15.78 44.08run 14 8.43 4.54 2.36 10.35 0.01 0.00 0.92 1.55 28.17 17.14 45.31run 15 8.13 4.54 2.34 9.95 0.02 0.01 1.37 2.47 28.81 18.51 47.33run 16 7.40 4.71 2.34 9.30 0.03 0.01 2.20 7.00 32.99 20.16 53.15

122

Table F3. Summary of residual standard deviations for model weighting runs

Resid Std Dev Resid Std Dev

Run NEFSC MRFSS SEAMAP VIMS COMM RECScrap/Disc ShrimpBase 0.607 0.450 0.475 0.647 0.067 0.023 0.124 0.181Index x 2 0.586 0.441 0.451 0.584 0.084 0.032 0.172 0.194Original 0.586 0.475 0.442 0.578 0.085 0.035 0.192 0.187run 1 0.511 0.485 0.432 0.596 0.037 0.024 0.335 0.456run 2 0.499 0.485 0.429 0.574 0.042 0.028 0.357 0.567run 3 0.488 0.483 0.425 0.549 0.049 0.033 0.389 0.690run 4 0.460 0.506 0.425 0.513 0.066 0.044 0.744 0.736run 5 0.522 0.480 0.431 0.604 0.029 0.019 0.284 0.351run 6 0.508 0.482 0.429 0.586 0.035 0.023 0.315 0.457run 7 0.494 0.483 0.426 0.561 0.041 0.028 0.340 0.603run 8 0.482 0.482 0.423 0.531 0.050 0.034 0.385 0.759run 9 0.536 0.473 0.430 0.607 0.022 0.015 0.218 0.264run 10 0.525 0.473 0.428 0.594 0.026 0.017 0.251 0.325run 11 0.506 0.477 0.426 0.575 0.032 0.022 0.293 0.470run 12 0.491 0.480 0.424 0.546 0.041 0.027 0.327 0.647run 13 0.545 0.468 0.430 0.609 0.016 0.010 0.153 0.201run 14 0.539 0.465 0.426 0.597 0.018 0.012 0.177 0.227run 15 0.529 0.465 0.424 0.586 0.022 0.015 0.215 0.287run 16 0.505 0.474 0.425 0.566 0.030 0.020 0.273 0.482

123

Table F4. Likelihood estimates for individual indices and fleets under different recruitment-deviation weights. Model runs 1 to 6 represent all fleets and indices with λ =1. Model runs 7-12 represent the ‘original’ index weighting scheme where all the fleets and MRFSS index had λ =1 and the fishery independent indices had λ=2. Likelihood Individual Index Likelihood Individual Fleet

Run Rec_Dev Wt NEFSC MRFSS SEAMAP VIMS COMM REC Scrap/Disc Shrimp1 1 10.69 4.25 2.93 12.14 0.19 0.02 0.45 1.002 1.39 11.07 4.28 3.14 13.49 0.27 0.03 0.64 1.383 3.37 12.26 4.11 3.94 18.06 0.61 0.04 1.55 3.034 4.33 12.56 4.03 4.21 19.48 0.72 0.05 1.92 3.655 5.8 12.99 3.88 4.54 21.19 0.81 0.05 2.18 4.466 8.25 13.47 3.71 4.92 23.17 0.91 0.06 2.54 5.457 1 9.95 4.75 2.55 9.68 0.27 0.04 1.09 1.048 1.39 10.01 4.92 2.65 10.50 0.39 0.06 1.58 1.339 3.37 10.73 4.93 3.13 13.74 0.79 0.09 2.54 2.96

10 4.33 11.03 4.81 3.33 14.99 0.91 0.09 2.71 3.6911 5.8 11.39 4.63 3.59 16.58 1.05 0.10 2.97 4.6512 8.25 11.84 4.40 3.94 18.57 1.22 0.11 3.33 5.88

124

Table F5. Total likelihood components for indices fleets and recruitment deviations different recruitment-deviation weights. Model runs 1 to 6 represent all fleets and indices with λ =1. Model runs 7-12 represent the ‘original’ index weighting scheme where all the fleets and MRFSS index had λ =1 and the fishery independent indices had λ=2. Weightings Un-weighted Likelihood

Run Rec_Dev NEFSCMRFSSSEAMAP VIMSCOMM RECScrap

/Discard Shrimp Rec Dev Fleet Index Total1 1 1 1 1 1 1 1 1 1 9.60 1.66 30.00 41.272 1.39 1 1 1 1 1 1 1 1 7.37 2.32 31.98 41.673 3.37 1 1 1 1 1 1 1 1 3.02 5.24 38.38 46.634 4.33 1 1 1 1 1 1 1 1 2.23 6.34 40.29 48.855 5.8 1 1 1 1 1 1 1 1 1.53 7.51 42.60 51.646 8.25 1 1 1 1 1 1 1 1 0.93 8.95 45.27 55.157 1 2 1 2 2 1 1 1 1 13.31 2.45 26.92 42.688 1.39 2 1 2 2 1 1 1 1 10.72 3.35 28.09 42.169 3.37 2 1 2 2 1 1 1 1 5.21 6.38 32.52 44.11

10 4.33 2 1 2 2 1 1 1 1 4.06 7.41 34.15 45.6211 5.8 2 1 2 2 1 1 1 1 2.93 8.77 36.19 47.8912 8.25 2 1 2 2 1 1 1 1 1.90 10.55 38.75 51.20

125

Figure F1. Comparison of standard deviation of residuals to weight applied to the NEFSC trawl index Figure F2. Comparison of standard deviation of residuals to weight applied to the MRFSS index Figure F3. Comparison of standard deviation of residuals to weight applied to the SEAMAP index

NEFSC

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0 2 4 6 8 10Weighting

Std

De

v

MRFSS

0.43

0.44

0.45

0.46

0.47

0.48

0.49

0.50

0.51

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Weighting

Std

Dev

SEAMAP

0.42

0.43

0.44

0.45

0.46

0.47

0.48

0 2 4 6 8 10

Weighting

Std

De

v

126

Figure F4. Comparison of standard deviation of residuals to weight applied to the VIMS index. Figure F5. Comparison of standard deviation of residuals to weight applied to the Commercial fleet.

Commercial Fleet

0.00

0.02

0.04

0.06

0.08

0.10

0 2 4 6 8 10

Weighting

Std

Dev

VIMS

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0 1 2 3 4 5 6 7

Weighting

Std

De

v

127

Figure F6. Comparison of standard deviation of residuals to weight applied to the recreational fleet. Figure F7. Comparison of standard deviation of residuals to weight applied to the scrap/discard fleet.

Recreational Fleet

0.000.010.010.020.020.030.030.040.040.050.05

0 1 2 3 4 5 6

Weighting

Std

De

v

Scrap/Discard Fleet

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.0 0.5 1.0 1.5 2.0Weighting

Std

Dev

128

Figure F8. Comparison of standard deviation of residuals to weight applied to the shrimp bycatch fleet. Figure F9. The influence of weighting of the NEFSC index on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the fleets

Shrimp Bycatch

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.0 0.2 0.4 0.6 0.8 1.0 1.2Weighting

Std

Dev

NEFSC

0

5

10

15

20

25

30

35

40

0 2 4 6 8 10

Weighting

Lik

elih

oo

d

Fleet Rec Dev Index

129

MRFSS

0

5

10

15

20

25

30

35

40

0.5 1.0 1.5 2.0 2.5 3.0 3.5

Weighting

Lik

elih

oo

d

Fleet Rec Dev Index

Figure F10. The influence of weighting of the MRFSS index on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the fleets.

SEAMAP

0

5

10

15

20

25

30

35

40

0 2 4 6 8 10

Weighting

Lik

elih

oo

d

Fleet Rec Dev Index

Figure F11. The influence of weighting of the SEAMAP index on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the fleets.

130

VIMS

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6 7

We i ght i ng

Fleet Rec Dev Index

Figure F12. The influence of weighting of the VIMS index on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the fleets.

Commercial

0

5

10

15

20

25

30

35

40

0 2 4 6 8 10

Weighting

Lik

elih

oo

d

Fleet Rec Dev Index

Figure F13. The influence of weighting of the commercial fleet on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the indices.

131

Recreational

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6

W eig ht ing

Lik

elih

oo

d

Fleet Rec Dev Index

Figure F14. The influence of weighting of the recreational fleet on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the indices.

Scrap/Discards

0

5

10

15

20

25

30

35

40

0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

W eig ht ing

Lik

elih

oo

d

Fleet Rec Dev Index

Figure F15. The influence of weighting of the scrap/discards on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the indices.

132

Shrimp Bycatch

0

5

10

15

20

25

30

35

40

0.4 0.6 0.8 1.0 1.2

Weighting

Lik

elih

oo

d

Fleet Rec Dev Index

Figure F16. The influence of weighting of the shrimp bycatch on the likelihood components. Fleet= total likelihood of the fleets. Rec Dev = recruitment deviation component. Index= total likelihood of all indices. The dotted line represents a trend line for the recruitment deviations and the solid line the trend line for the indices. Figure F17. Total likelihood and major likelihood components across model runs.

0

10

20

30

40

50

60

70

80

Bas

e

Inde

x x

2

Orig

inal

run

13

run

9

run

14

run

5

run

10

run

15

run

1

run

6

run

11

run

16

run

2

run

7

run

12

run

3

run

8

run

4

Lik

elih

oo

d

Fleet Rec Dev Total Index

133

Figure F18. The cumulative proportion of 5,000 simulations relative to the total un-weighted likelihood of each run. The solid line indicates the median likelihood of the 5000 simulations (55.07).

Figure F19. The association between the standard deviation of the residuals and weighting (λ ) for the NEFSC index based on 5,000 simulations.

0.0

0.2

0.4

0.6

0.8

1.0

25 75 125 175Total un-weighted Likelhood

Pro

po

rtio

n o

f ru

ns

NEFSC

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 5 10 15 20Weighting

Std

Dev

Res

id

134

Figure F20. The association between the standard deviation of the residuals and weighting (λ ) for the SEAMAP index based on 5,000 simulations. . Figure F21. The association between the standard deviation of the residuals and weighting (λ ) for the VIMS index based on 5,000 simulations

SEAMAP

0.0

0.2

0.4

0.6

0.8

1.0

0 5 10 15 20Weighting

Std

Dev

Res

id

VIMS

0.00.20.40.60.81.01.21.4

0 5 10 15 20Weighting

Std

Dev

Res

id

135

Figure F22. The association between the standard deviation of the residuals and weighting (λ) for the MRFSS index based on 5,000 simulations.

Shrimp Bycatch

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 5 10 15 20Weighting

Std

Dev

Res

id

Figure F23. The association between the standard deviation of the residuals and weighting (λ) for the shrimp bycatch based on 5,000 simulations

MRFSS

0.0

0.2

0.4

0.6

0.8

1.0

0 5 10 15 20Weighting

Std

Dev

Res

id

136

Figure F24. The association between the standard deviation of the residuals and weighting (λ) for the scrap/discards based on 5,000 simulations. Figure F25. The association between the standard deviation of the residuals and weighting (λ) for the commercial landings based on 5,000 simulations.

Scrap/Discards

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 5 10 15 20Weighting

Std

Dev

Res

id

Commercial

0.00.5

1.01.5

2.02.5

3.03.5

0 5 10 15 20Weighting

Std

Dev

Res

id

137

Figure F26. The association between the standard deviation of the residuals and weighting (λ) for the recreational landings based on 5,000 simulations.

Recreational

0.0

0.51.0

1.52.0

2.53.0

3.5

0 5 10 15 20Weighting

Std

Dev

Res

id

138

Appendix G. A re-assessment of the status of the Atlantic croaker population in the Mid-Atlantic (New York to North Carolina). The recent review of the Atlantic States Marine Fisheries Commission (ASMFC) Atlantic croaker stock assessment identified seven areas of concern. The ASMFC Atlantic croaker management board prioritized the recommendations of the peer review panel and identified four issues that the Atlantic croaker technical committee (TC) needed to address immediately. The management board also made the decision to use an analytical model that only incorporated data from the mid-Atlantic (North Carolina to New York) but requested the TC to evaluate the basis for this at a later date (Issue 5 – review panel recommendations). This report does not address developing alternate modeling approaches such as the Collie-Sissenwine catch survey and delay-difference models (Issue 6 of review panel recommendations). We summarize the changes made to the age structured production model and the results from the revised version. Summary of Changes In this revision of the model the following changes were made:

1) Estimates of North Carolina and Virginia’s scrap landings were included in the model. A model where scrap estimates were treated as a separate component was chosen over one where scrap landings were included as part of the commercial landings.

2) Using data from the NEFSC observer database, estimates of at-sea discards for the gill net and otter trawl fishery have been included.

3) The NEFSC trawl survey index has been extended to the entire time series, and the stratified mean estimates in numbers were used.

4) The VIMS spring index has been included in the model. 5) The model now estimates initial SSB: SBB virgin ratio. 6) The selectivity patterns used for the fleets has been refined using selectivity patterns

estimated from an ‘un-tuned’ separable VPA by incorporating the length and age data for Virginia’s and North Carolina’s commercial fishery (1989-2002) and the recreational fishery’s size distribution (1981-2002).

7) Commercial landings for 2002 were updated.

139

Data changes

Harvest and discards The major changes in landings data were the inclusion of scrap estimates from North Carolina and Virginia and at-sea discards from the gill net and trawl fishery. The technical committee also evaluated the potential inclusion of shrimp bycatch estimates, but decided the data quality was poor and had little confidence in the estimates. However, the implications of not including estimates of Atlantic croaker bycatch in the shrimp fishery were evaluated through sensitivity analysis. For the scrap landings, the model includes estimates developed by the North Carolina Division of Marine Fisheries (NCDMF) of Atlantic croaker scrap (bait) for 1986 to 2002 (see Appendix A for detailed report). For estimates of North Carolina’s scrap landings from 1973-1985, the technical committee evaluated estimates created using a variety of methods and determined the most appropriate estimates were those based on the average ratio of scrap to unclassified finfish landings from the NMFS database between 1986-1990 (see Appendix A for details). In the model, scrap landings and discards have been combined. It was assumed that the selectivity pattern and size distribution of the discards would be similar to the scrap landings. Commercial landings for 2002 were updated from the NMFS database. Landings estimates used in the revised model are presented in Table G1. Between 1973 and 2002 the relationship the different sources of removals has changed. From 1973 and 1995 scrap/discards accounted for an average 20% of the annual removals (ranged between 14-30 %). From 1996-2002, scrap/discards accounted for an average 3% of the removals. Estimates of scrap/discards reached their peak in 1979 (3,200 MT) and since then shown declined to their lowest levels in 2002 (425 MT). Between 1973 and 1995, scrap/discard removals averaged 1,687 MT per year, whereas between 1996-2002 scrap/discards averaged 595 MT per year.

Indices In the revised model, data from the NEFSC trawl survey was re-examined (see Appendix E for details). Estimates used in the model are based on the stratified means in numbers from 1973 to 2002 (Appendix E). In addition, the Virginia Institute of Marine Science (VIMS) spring trawl index was included in the revised model. This index is a young-of -year (YOY) index and estimates are the geometrical mean in numbers. While this index is spatially limited to Chesapeake Bay, it extends across the time series (1973-2002). Preliminary analysis revealed that the pattern in recruitment deviations were closely associated with indices that had a strong age 0 component. The TC concluded that including the VIMS index into the revised model was beneficial, in that recruitment deviations would be more closely associated with an additional index and would also help in the estimation of parameters in the stock-recruit relationship. Preliminary analyses revealed that unless

140

the model included abundance indices that covered the early part of the time series (~1973), the initial SSB: SSB virgin ratio was poorly estimated. Indices used in the revised model are presented in Table G2. Model changes The technical committee considered the options available for incorporating the scrap/discard landings into the model. The two options were:

1) Including the scrap landings as a “pseudo fleet” with its unique selectivity pattern.

2) Incorporating the scrap landings into the commercial landings and adjusting the selectivity pattern accordingly.

The technical committee evaluated the advantages and disadvantages of the two model configurations. For the two fleet model, where scrap/discards were included with the commercial landings, advantages include a model with fewer parameters. Including the scrap/discards with the commercial fleet is intuitively pleasing, as it is a component of the commercial landings. However, the scrap/discards tend to be small fish, with a different selectivity pattern to that of the landings (see selectivity section). In addition, there was evidence that the relationship between scrap and landings has not been constant across the time series. This was clearly evident for data from North Carolina. As such, the TC concluded treating the scrap/discards as a ‘pseudo fleet’ was the most appropriate way to include the data in the model. The two model configurations produced similar parameter estimates and fishing mortality rates for the recent time series. However, estimates of the fully selected fishing mortality rates for the commercial fishery in the two fleet model, reached the upper bound for some years in the early part of the time series. To account for different selectivity patterns among the fleets, the fishing mortality rates are expressed as the average fishing mortality rate from ages 1 to 10+, weighted by population size. In addition, the revised model estimates the initial SSB: SSB Virgin ratio. In the original version, the initial SSB:SSB virgin ratio was set to 0.75 and included as a term in the sensitivity runs. For the base model, the steepness (0.76), natural mortality (0.3), growth, and length-weight relationships used were similar to those in the original version.

Selectivity An important deterministic component of the age structured production model is the selectivity pattern used for each of the fleets and indices. As this re-analysis included estimates of at-sea discards and the scrap landings from Virginia and North Carolina, which comprise of the majority of landings, the available data from these two states were re-evaluated and used to describe the selectivity components of the fleets.

141

Data and methods for estimating selectivity of fleets For North Carolina, landings of the marketable Atlantic croaker and scrap by gear type and year in numbers had been estimated by the NCDMF (NCDMF 2003). In addition, the NCDMF had also estimated the length composition of the marketable landings and scrap by gear type from 1986 to 2002 (NCDMF 2003). For Virginia, the Virginia Marine Resources Commission (VMRC) collected length-weight data characterizing the commercial fishery from 1989 to 2002 by gear and market category. In addition, landings by gear type and market category (in pounds) were also available from the VMRC from 1989-2002. Landings in numbers and size-class were estimated from 1989-2002 by gear and market category using the VMRC bio-profile data on length and weight (in 20 mm size classes). Using the length-weight data for each market and gear category, the landings in numbers were then estimated using the size-weight relationship. For Virginia’s scrap component, the truncated size distribution (all fish less than 8 inches and 50% of fish of 9 inches TL) was used to estimate the scrap component (see section on estimating Virginia Scrap landings). For the recreational fishery, an annual weighted size distribution had been developed for the mid-Atlantic from 1981 –2002 (see section 5.2.1.6 in original report). As North Carolina and Virginia account for more than 90 % of the commercial landings of Atlantic croaker, using a catch-at-age matrix based on the data from these states would provide a reasonable description of selectivity patterns for the commercial fishery. Using the North Carolina otolith age database, an age-length key was developed combining data from all years and gear types in 20 mm intervals. Using the available bio-profile data, a set of catch-at-age matrices for the Virginia’s and North Carolina’s market and scrap landings were developed for the period 1989 to 2002 and the mid-Atlantic recreational landings from 1981-2002. In developing the catch at age matrices length classes of 20 mm were used as some of the length data was in 20mm increments. Given the lack of length data prior to the mid-1980’s it was assumed that the selectivity pattern based on the period 1989-2002 would be applicable to the early part of the time series (1973-1988). Three composite catch-at-age matrices were developed for the commercial landings.

1) An annual market grade catch-at-age matrix for landings from North Carolina and Virginia was developed (1989-2002).

2) An annual scrap grade catch-at-age matrix for estimates from North Carolina and Virginia was developed (1989-2002).

3) Market and scrap catch-at-age matrices from North Carolina and Virginia were combined to produce an annual catch-at-age matrix for scrap and market grades combined (1989-2002).

142

Catch-at-age matrices for the commercial, scrap, and recreational fishery are shown in Tables G3 to G5. An ‘un-tuned’ separable VPA (Clay 1990) was used to estimate the selectivity patterns for each of the three composite commercial catch at age matrices and the recreational catch at age matrix. The separable VPA uses Pope and Shepard’s (1982) method for estimating selectivity. Input parameters for the model include a catch-at-age matrix, an estimate of a fully selected age, shape of selectivity curve (dome or flat), and terminal fishing mortality rate. In estimating selectivity patterns for the different components, a series of terminal fishing mortality rates between 0.1-0.3 were used. Selectivity patterns were insensitive to the choice of terminal fishing mortality rate. For the combined market and scrap catch-at age-matrix, market catch-at-age matrix and recreational catch-at-age, a flat-topped selectivity pattern was used where the ascending limb was estimated. For these components, a fully selected age of four was used. For the scrap catch-at-age component, a dome shaped selectivity pattern was used, and a fully selected age of one was used. Based on the model output, choices used for the fully selected ages for each of the components appeared to be appropriate.

The selectivity patterns for the four fleet components are presented in Table G6. The selectivity patterns for the MRFSS and NEFSC indices were also revised. For the NEFSC index, the age composition of the index (see Appendix E) was used to revise the original selectivity pattern. The selectivity pattern for the MRFSS index was used to model the recreational fleets. Since the VIMS index is an age 0 index, age 0 was considered fully selected (1), and all other ages were set at 0 (Table G7).

Output/results

Goodness of fit of model used

The goodness of fit of statistical model is judged by how well the predicted estimates match the observed estimates. In general, the base model appeared to fit the data well, with few outliers. Examination of the standardized residuals for each of the fleet components indicates few data points exceeded an absolute value of 2.0 (Figure G1). However, for the commercial and recreational fleets, predicted estimates slightly underestimated landings in recent years. For the indices, standardized residuals also indicated few outliers (absolute values > 2), and there appeared to be little sign of serial correlation in the error terms. However, for the NEFSC trawl survey, while the model captured the general trend of the index, it poorly fitted the peak estimates of 1975, 1994, 1999, and 2002. For these years, estimates were on average two times greater than estimates in adjacent years (Table G2). The predicted estimates of the other indices also captured the general trends of their respective indices adequately. However, as with the NEFSC trawl index, peak estimates for 1999 in the MRFSS, 1992 in the SEAMAP, and 1983-86 in the VIMS index were underestimated. For the period

143

1983-86 the VIMS index indicated higher than average estimates, whereas the NEFSC and MRFSS indices indicated relatively low estimates during this period.

Parameter estimates

In the revised model, 125 parameters were estimated. The estimated parameters include the initial SSB: SSB virgin ratio, the number of virgin recruits (R0), a catchability coefficient for each of the indices (4), an annual recruitment deviation from the stock-recruit relationship for 1974-2002 (29), and a fully selected fishing mortality rate for each of the fleets (90). For the base model estimated R0 at 170 million fish, the initial SSB: SSB virgin ratio at 0.296, the catchability coefficient for the NEFSC index at 6.53778E-07, the catchability coefficient of the MRFSS index at 2.71784E-09, the catchability coefficient of the SEAMAP index at 2.54743E-06 and the catchability coefficient of the VIMS index at 6.71846E-09. Estimates of the fully recruited fishing mortality estimates and recruitment deviations from the base model are presented in Table G8.

Exploitation rates In the revised model, fishing mortality rates (F) are based on the average population weighted F for ages 1-10+. Exploitation rates (u) are expressed as the predicted catch (in numbers)/ population estimate (in numbers). Unless, otherwise noted fishing mortality rates referred to in the text are to the average instantaneous fishing mortality rate. Fishing mortality rates for Atlantic croaker exhibit a cyclical trend over the time series. From 1977 to 1979, F rose rapidly reaching a maximum of 0.5 in 1979. From 1980 onwards, F rapidly declined reaching its lowest levels in 1992 (Figure G3; Table G8). Since 1993, F has gradually increased and between 1997 and 2002 remained relatively stable at around 0.11 (Figure G3; Table G8). Exploitation rates followed a similar trend to F, reaching it’s maximum in 1979 (u=0.25; Table G8). Exploitation rates in recent years (1997-2002) have been low ranging between 0.05 and 0.08 (Table G8).

Abundance estimates For the base mid-Atlantic run, the trend in population abundance indicates a step-wise increase reaching a peak of 974 million fish in 1999 (Table G9). Population estimates from 1999 to 2002 have ranged from 663 to 974 million fish. The number of age 0 fish in the population exhibited a series of periodic spikes in 1975, 1983, 1991, 1998, and 2002 (Figure G4; Table G9). Between 1999 and 2002 the number of age 0 fish have ranged between 100-375 million fish. Spawning stock biomass estimates (estimated as the proportion of mature females) exhibit a cyclical trend over the time series. From the early 1970’s to 1983 spawning stock biomass declined to its lowest level (11,746 MT). Since 1984, spawning stock biomass has increased in three distinct phases, with estimates reaching a maximum in 1996 (Figure G4, Table G9). Between 1999 and 2002

144

spawning stock biomass estimates have ranged between 80-91,000 metric tons (Table G9).

Precision of parameter estimates Burnham and Anderson (1998) define precision as “ a property of an estimator related to the amount of variation among estimates from repeated samples”. The model developed in excel, does not provide any estimates of precision. For models run using AD model builder, estimates of standard deviation are based on the delta method, which approximate the variance estimates. Variance estimates using the delta method are biased to the lower range of the spectrum when additional constraints are imposed on the model (ASMFC 2003). Confidence bounds on the parameters can be estimated using bootstrap procedures. However, the estimates derived are likely to be biased (Hilborn and Walters 1992). Ideally, the relative levels of confidence of the parameter estimates should be evaluated using methodology such as the “operating model concept” described in Hilborn and Walters (1992) or Bayesian methods. These are part of the long-term objectives in the model’s development. Examination of alternate weighting strategies of the likelihood components revealed that selecting a weighting profile using an objective set of criteria were difficult (see Appendix F). However, the influence of alternative weighting criteria on selected performance statistics may be a useful method to characterize the uncertainty around those estimates. For the base model, we ran a simulation of 3,500 runs, where a random and independent weight (λ) ranging between 0-20 was selected from a uniform distribution for each of the fleet and index terms. The performance statistics evaluated were the average fishing mortality rate per year (ages 1-10+), spawning stock biomass estimates per year, and the ratios of average fishing mortality rate to Fmsy and spawning stock biomass to SSBmsy in 2002. Average fishing mortality rates from 1973–2002 from the simulation were consistent with patterns observed for the base model. The inter quartile range (25-75th percentile) for F2002 from the simulations ranged from 0.015 to 0.11 (Figure G5). For 2002, average fishing mortality rates from the base model was close to the 75th percentile of the simulation runs (Figure G5a) (average F=0.11). Spawning stock biomass trends from the simulation runs also show a similar trend to estimates derived from the base run (Figure G6). The inter quartile range for spawning stock biomass estimates from the simulation in 2002 ranged between 71,000 and 120,000 MT. In comparison, estimates of spawning stock biomass in 2002 from the base model was 80,000 MT, close to the value of 25th percentile of the simulation runs. For both fishing mortality and spawning stock estimates, estimates determined from the base run appear to be more pessimistic (conservative) when compared to other potential weighting schemes. This assumes that 3,500 simulations capture a wide range of weightings.

145

Sensitivity analysis In the original model, the TC identified five deterministic inputs that the parameter estimates were likely to be sensitive to. These parameters were in the initial SSB:SSB virgin ratio, selectivity patterns for the early age classes for the commercial and recreational fishery, steepness, and natural mortality. In the revised version, the initial SSB: SSB virgin ratio is estimated and the selectivity patterns for the fleets are based on estimates derived from an ‘un-tuned’ separable VPA. As such, sensitivity analysis for the revised model examined the effects of varying steepness and natural mortality estimates on the parameter estimates and biological reference points. Choice of steepness estimates can have a large impact on stock status. The TC identified a subjective weighting for a range of natural mortality estimates (see section 6.1 in original document). These weightings were used to create a probability distribution for natural mortality. For steepness, the prior distribution developed by Myers et al. (2002) was used (see section 7.4 of original report for distributions). Steepness and natural mortality estimates were selected from the probability distributions and the model was run 2,500 times. While this method, does not capture all of the uncertainty associated with the model, it does capture two of the major sources of uncertainty based on an assigned distribution for each of the parameters. Examination of the likelihood profile of the 2,500 runs shows a strong correlation between the total likelihood and steepness (Pearson Corr. =-0.7; Figure G7). The best-fitting models were associated with steepness estimates ~1. There appears to be little correlation between the total likelihood and natural mortality estimates used (Pearson Corr =0.3). Fishing mortality rates from the sensitivity runs indicate that estimates up to the early 1980’s were associated with a high degree of variability. Fishing mortality estimates from recent years have been relatively stable and show low variability across runs (Figure G8; Table G10). For 2002, the inter quartile range of fishing mortality estimates were between 0.08 and 0.12. F2002 from the base run was 0.11. Spawning stock biomass estimates from the sensitivity runs are presented in Table G11. Trends in spawning stock biomass under varying steepness and natural mortality rates show greater variability in recent years than those for the early part of the time series (Figure G9). For 2002, the inter quartile range of spawning stock biomass estimates was 80,000 and 110,000 Metric tons. Spawning stock biomass estimates for 2002 from the base run was 80,000 MT. Based on the sensitivity runs, it appears that ~25% of the runs had higher fishing mortality estimates than those for the base run and ~25% of the sensitivity runs had spawning stock biomass estimates lower than the base run.

Biological reference points As part of the model configuration, a Beverton and Holt stock recruitment relationship re-parameterized in terms of steepness is included. Estimates of the virgin recruitment for the base mid-Atlantic were 169 million fish. The stock recruitment curves for the base mid-Atlantic are presented in Figure 10. For the base mid-Atlantic model a wide scatter between recruits and spawning stock was evident, and estimates for the recent

146

part of the time series are scattered around the region where the replacement line meets the stock-recruit curve (Figure G10).

Overfishing definition The benchmarks for the mid-Atlantic region are:

F threshold - Fmsy Biomass threshold - 0.7 SSBmsy F target – 0.75 Fmsy Biomass target – SSBmsy

Estimates of Fmsy from the base mid-Atlantic model was 0.39 and SSBmsy was equal to 28,932 MT. Estimates of average fishing mortality rates from the base mid-Atlantic model of 0.11 indicate that 2002 estimates were below the target and threshold levels (Figure G11). Estimates of SSB from the base mid-Atlantic model relative to the proposed target and threshold SSB levels are shown in Figure G12. Recent estimates of SSB (~80,000 MT) are above both the proposed target and threshold levels. For 2002, F:Fmsy ratio was 0.263 and SSB:SSBmsy ratio 2.78. Uncertainty in the estimates of the current status of the stock (in 2002) were examined at three levels; 1) the sensitivity of the base model to alternate weightings of the likelihood components; 2) sensitivity of the model to alternative steepness and natural mortality estimates; and 3) the implications of not including shrimp bycatch estimates. Based on the base run’s sensitivity to weighting of the likelihood components, and the sensitivity of the model to alternate steepness and natural mortality estimates, estimates derived from the base run appear robust. From the sensitivity analysis on weighting of the likelihood terms, 90 % of the simulations had F2002:Fmsy ratios less than 0.44 (Table G12; Figure G13a). Biomass reference points from the weighting analysis indicated that 10% of the runs had SSB2002: SSBmsy ratios less than 2.27 (Table G12; Figure G13b). Model sensitivity to steepness and natural mortality estimates also indicated the stock was most likely below the fishing mortality targets and thresholds and above the biomass targets and thresholds; 90 % of the simulations had F2002:Fmsy ratios less than 0.44 (Table G12; Figure G14a) and 10% of the runs had SSB2002: SSBmsy ratios less than 2.16 (Table G12; Figure G14b). The TC discussed the quality of estimates of Atlantic croaker discards from the shrimp fishery in depth. The TC concluded further work needed to be carried out on estimating Atlantic croaker bycatch in the shrimp fishery. Evaluating the effectiveness of developing estimates of discards by combining available information on the effectiveness of bycatch reduction devices with estimates of the effective ‘swept area’ by the shrimp fishery and abundance estimates from the SEAMAP indices need to be explored in more detail. The shrimp fishery has undergone significant changes in efficiency with the introduction of bycatch reduction devices and turtle excluder

147

devices. Based on the available length data, the majority of Atlantic croaker caught as shrimp bycatch were age 0 fish. The preliminary estimates of Atlantic croaker bycatch may not capture the inter-annual variability across the time series, as estimates for 1973-1991 are based on 39 tows, 1992-1998 on 685 tows, and 1999-2002 on 56 tows (See appendix C for details). While the uncertainty surrounding the estimates across years is high, estimates for 1994 are likely to be good since they were based on 522 tows (67% of the available data). Estimates of Atlantic croaker from the shrimp fishery in 1994 were 5,200 MT of age 0 fish. Given the potential magnitude of estimates known with reasonable confidence, sensitivity of the biological reference points to the inclusion/non-inclusion of estimates from shrimp bycatch was examined. This analysis assumes that the preliminary estimates of Atlantic croaker bycatch are the best available and that the standard error estimates associated with the three ratios used, capture the inter-annual variability. For the analysis, the estimates of Atlantic croaker from the shrimp bycatch were treated as a separate fleet with a selectivity pattern that treated all age 0 fish as fully selected (1) and all other age classes as not being selected (0). Average fishing mortality estimates evaluated were age 0-10+ weighted by the population. Using the standard error estimates for the three ratio estimators, annual estimates of Atlantic croaker from the shrimp fishery were determined using:

FisheryShrimpfromLandACSEratiodevnormrandratioShrimpACShrimpLandEstimate : The simulation was run 1,000 times and the range of Atlantic croaker bycatch estimates evaluated are shown in Figure G15. Estimates of Fmsy, SSBmsy, and the current status of the stock were used as performance statistics (Table G13). Average fishing mortality rates (ages 0-10+) in 2002 ranged from 0.06 to 0.176 with 50% of the simulations having values less than 0.105 (Figure G16a). Spawning stock biomass estimates in 2002 from the simulation runs ranged from 77,000 to 149,000 MT with 50% of the values being less than 111,388 MT (Figure G16b). In comparison, the average fishing mortality rate from the base run in 2002 was 0.11 (ages 1-10+) and the spawning stock biomass estimate in 2002 was 80,328 MT. When including estimates of Atlantic croaker caught as shrimp bycatch, simulations revealed that the current status of the stock was similar to the base run where shrimp bycatch were not included; the stock is not overfished or undergoing overfishing. However, biomass reference points from the simulation runs indicated higher SSBmsy values and the lower estimates of SSB2002:SSBmsy than those obtained for the base model. The range of estimates for Fmsy (~0.4; Figure G17a) was similar to the base model (~ 0.39). SSBmsy estimates from the simulation (ranged from 48,000-67,000 MT with a median of 56,467 MT; Table 13; Figure G17b) and were much higher than those for the base run (28,932 MT). Differences in Spawning stock biomass estimates are most likely a result of the model accounting for the increased removals as part of the shrimp bycatch by increasing the population estimates. The ratios of F2002:Fmsy and SSB2002:SSBmsy for the sensitivity runs are summarized in Table G13. The ratio of F2002:Fmsy ranged from 0.14-0.43 with 50% of the runs having estimates below 0.26 (Figure G18a). In comparison F2002:Fmsy from the base model was 0.263 (based on ages 1-10+). The ratio of SSB2002:SSBmsy for the simulations ranged from 1.55 to 2.27, with

148

50% of the runs having estimates less than 1.98 (Figure G18b). In comparison, the SSB2002:SSBmsy ratio for the base run was 2.78. Recommendations and findings The mid-Atlantic model, which is the core of the population, indicates fishing mortality rates were high in the mid 1970’s, abruptly declined, and has been low and stable since the mid 1990’s. Between 1973 and 2002 the relationship between the different sources of removals has changed. In particular, estimates of scrap/discards reached their peak in 1979 (3,200 MT) and since then declined to their lowest levels in 2002 (425 MT). Between 1973 and 1995, scrap/discard removals averaged 1,687 MT per year, whereas between 1996-2002 scrap/discards averaged 595 MT per year. It appears that the significant reduction in removals of predominantly age 1 and younger fish may have contributed to relatively stable fishing mortality and spawning stock biomass estimates since the mid 1990’s. In relation to the proposed reference points the Atlantic croaker population is not overfished or undergoing overfishing. The commercial and recreational catch-at-age data from recent years also shows an increasing age distribution, with a few fish of 12 years being observed in the commercial landings. Anecdotal evidence from the mid-Atlantic indicates an expansion of the population at the northern part of the range. For example, in Delaware, fishery independent indices indicate a recent increase in abundance of Atlantic croaker in the region (D. Kahn, personal communication). In addition, both commercial and recreational landings from New Jersey and Delaware have increased recently. The population has benefited from good recruitment in recent years, which may also be tied to the regulatory changes that have affected some of the fisheries that indirectly target Atlantic croaker (see Section 3.2 of original report). While this analysis does not capture all of the sources of uncertainty, examination of the effects of alternate weightings of the likelihood components and alternate steepness and natural mortality estimates indicate that reference points derived from the base run are relatively robust. The reference points suggest that there was less than a 10% chance that the population is overfished or undergoing overfishing. Sensitivity analysis evaluating the inclusion/non-inclusion of shrimp bycatch estimates, indicate that SSBmsy estimates are sensitive to the inclusion of Atlantic croaker caught as shrimp bycatch. However, increased SSBmsy estimates are also accompanied by higher SSB estimates. The ratio of SSB2002:SSBmsy when shrimp bycatch is included indicates that the stock is unlikely to be below the threshold estimates. Of concern, would be management goals that define biomass reference points in absolute terms. There appears to be some justification for revising the reference points for the biomass target and threshold to relative terms until a more comprehensive evaluation of Atlantic croaker from shrimp bycatch can be carried out.

149

Literature cited ASMFC.2003. Atlantic menhaden 2003 stock assessment report. Atlantic States Marine Fisheries Commission. Burnham, K. P. and R.D. Anderson. 1998. Model selection and inference. A practical information-theoretic approach. Springer-Verlag. New York. Clay, D. 1990. TUNE: a series of fish stock assessment computer programs written in FORTRAN for microcomputers (MS DOS). Internat. Comm. Conserv. Atl. Tunas, Coll. Vol. Sci. Pap. 32:443-460. Hilborn, R and C.J. Walters. 1992. Quantitative Fisheries Stock Assessment. Choice dynamics and uncertainty. Kluwer Academic Publishers. Myers, R.A., N.J. Barrowman, R. Hilborn, D.G. Kehler. 2002. Inferring Bayesian priors with limited directed data: Applications to risk analysis. North American Journal of Fisheries Management 22:351-364. North Carolina Division of Marine Fisheries (NCDMF). 2003 Assessment of North Carolina Commercial Finfisheries, 1997-2002. Final Performance report for Award Number NA 76 FI 0286, 1-3. North Carolina Department of Environment and Natural Resources. North Carolina Division of Marine Fisheries. Pope, J.G. and Shepard, J.G. 1982. A simple method for the consistent interpretation of catch-at-age data. Journal du Conseil International pour l’Exploration de la Mer 40:176-184.

150

Table G1. Landings estimates used in revised model (metric tons)

Year CommercialRecreationalScrap/Discards1973 2,611 1,027 1,316 1974 3,515 1,284 1,727 1975 7,484 2,325 1,631 1976 10,300 3,292 1,761 1977 13,506 3,547 2,236 1978 13,292 3,211 2,680 1979 10,385 2,036 3,193 1980 9,923 1,019 2,579 1981 5,289 449 1,790 1982 4,967 366 1,627 1983 3,357 432 1,693 1984 4,570 619 2,002 1985 4,955 546 1,702 1986 5,459 1,067 930 1987 4,756 880 1,705 1988 4,678 1,958 1,715 1989 3,628 938 1,664 1990 2,709 614 1,275 1991 1,651 1,004 1,019 1992 1,905 1,005 858 1993 4,017 1,375 952 1994 4,866 2,116 1,268 1995 6,309 1,713 1,484 1996 9,452 1,821 710 1997 12,231 3,460 753 1998 11,471 3,533 459 1999 12,113 3,134 715 2000 12,091 4,375 596 2001 12,970 4,955 511 2002 11,717 4,170 424

151

Table G2. Indices and estimates used in revised model

YEAR NEFSC MRFSSSEAMAP VIMS (Numbers)(Numbers) (Weight)(Numbers)

1973 38.07 0.121974 143.20 2.041975 638.21 2.631976 397.61 1.081977 119.35 0.151978 161.72 0.081979 15.64 2.181980 88.53 0.521981 31.77 0.235 0.071982 9.11 0.228 0.111983 231.94 0.674 6.591984 267.61 0.648 1.631985 213.97 0.397 4.981986 127.11 0.616 2.971987 111.96 0.690 4.241988 31.65 0.807 0.321989 99.64 0.860 16.35 0.601990 79.82 0.625 15.03 0.431991 260.53 0.899 79.44 4.411992 216.19 0.795 150.26 1.281993 140.88 0.957 26.54 2.171994 478.57 1.287 65.90 0.901995 189.36 0.855 60.84 1.061996 203.99 0.855 31.91 0.191997 159.14 1.232 10.19 1.471998 344.79 1.424 59.02 1.191999 734.45 2.108 87.86 1.502000 387.65 1.517 25.63 0.602001 177.64 1.371 21.73 0.362002 939.82 1.135 25.53 1.59

152

Table G3. Catch at age matrix for Atlantic croaker commercial landings from Virginia and North Carolina used to determine selectivity pattern

Year Age0 Age1 Age2 Age3 Age4 Age5 Age6 Age7 Age8 Age9 Age10+1989 1,913,8968,284,7184,878,4222,571,3101,060,693 308,536 153,695 45,575 20,674 5,194 4,8341990 3,001,5258,024,3424,093,3611,596,628 492,309 72,574 38,989 5,071 1,193 178 1981991 1,591,8195,681,5922,814,3731,039,527 316,941 43,376 23,218 2,047 481 50 491992 1,066,9815,443,8073,218,4121,347,402 428,387 68,260 33,679 5,271 1,273 223 2241993 1,462,6149,117,3846,716,7313,302,3961,131,509 222,095 102,459 26,749 6,617 1,205 1,2491994 1,537,2519,236,3107,036,7383,666,4621,362,637 338,939 174,281 82,920 23,638 6,706 4,8961995 1,191,3617,488,0016,369,8304,486,0402,280,852 874,727 480,024 321,302 95,570 25,996 21,4931996 544,3205,548,4226,801,2586,058,6763,774,5201,803,5041,096,841 877,463 265,582 83,821 57,2381997 463,7715,196,6386,513,1395,976,2573,796,7291,895,4391,179,833 980,822 300,052 104,244 64,2981998 326,0554,025,1995,838,9816,282,4364,763,8912,646,4591,713,5391,451,298 450,857 148,517 99,2551999 236,7403,283,2044,977,6175,630,5034,616,6052,845,1881,995,9071,913,074 586,656 214,108 135,8782000 298,9923,591,1525,078,6115,428,2594,420,4462,741,7901,974,8241,879,379 582,600 206,648 134,3842001 281,0313,236,2204,614,5245,156,5654,506,9463,001,9802,207,8302,251,971 692,983 269,526 152,0942002 191,6363,023,4544,733,4885,403,2684,306,6722,605,2321,778,1501,668,526 513,716 193,680 112,311

Table G4. Catch-at-age matrix for Atlantic croaker scrap from Virginia and North Carolina used to determine selectivity pattern.

Year Age0 Age1 Age2 Age3 Age4 Age5 Age6 Age7 Age8 Age9 Age10+

1,989 3,084,834 10,419,742 3,205,238 662,641 166,240 6,613 14,550 21 7 0 01,990 4,204,683 7,546,689 2,059,344 427,228 105,592 4,072 9,630 117 29 0 01,991 3,120,426 8,135,978 1,978,838 316,430 78,965 1,733 6,720 36 9 0 01,992 2,542,800 8,242,979 2,169,508 440,018 111,592 6,055 9,925 467 67 67 671,993 2,559,385 5,930,994 2,087,439 568,524 144,932 9,508 12,858 99 17 11 111,994 5,029,502 9,273,672 1,933,851 487,346 119,883 7,369 10,760 118 28 2 21,995 4,018,781 9,100,379 1,962,377 435,327 108,533 6,644 9,405 190 44 5 51,996 568,967 2,290,915 1,075,225 410,810 131,689 22,087 11,571 2,260 579 55 761,997 899,437 2,434,017 905,559 303,971 86,870 9,905 6,610 443 118 29 291,998 848,224 856,254 403,006 146,077 42,455 5,532 3,919 568 162 19 291,999 2,219,998 2,832,170 459,603 150,838 44,668 7,695 4,436 933 240 27 272,000 1,175,057 2,267,978 487,748 144,718 43,550 6,493 3,623 671 190 25 332,001 386,140 1,134,278 399,721 136,608 42,470 6,470 3,402 511 133 14 14

2,002 178,416 668,908 271,092 125,459 45,467 10,393 4,720 1,559 357 73 86

153

Table G5. Recreational catch-at-age matrix used to determine selectivity pattern.

Year Age 0 Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Age 7 Age 8 Age 9Age 10+

1,981 212,633 1,007,408 474,407 222,291 106,742 43,123 26,313 12,453 6,560 2,705 1,1571,982 99,066 433,047 194,312 105,260 53,425 22,923 14,939 14,778 3,752 1,495 8,8921,983 1,514,569 2,007,776 446,687 136,410 58,064 22,374 17,641 7,245 4,320 1,391 1,1081,984 448,335 2,288,610 1,124,814 479,917 193,841 57,935 37,124 11,969 5,941 885 1,7141,985 359,281 1,624,282 812,758 351,399 135,355 36,809 21,500 6,239 3,110 612 8121,986 651,207 3,430,795 1,841,266 774,127 289,262 79,239 49,136 16,132 9,808 3,420 1,7271,987 321,474 1,861,768 1,264,185 684,409 297,092 97,083 53,158 17,761 9,948 3,008 2,0311,988 518,828 2,718,699 1,767,141 1,211,467 688,538 302,130 175,665 74,724 41,137 13,777 8,3801,989 410,214 2,046,026 1,163,752 585,118 246,291 74,482 39,286 11,690 6,681 1,578 1,2201,990 590,312 2,085,115 863,083 319,217 134,549 41,594 26,248 8,398 5,300 837 8011,991 857,974 3,424,106 1,306,488 449,741 147,001 27,645 15,913 3,772 1,778 711 3511,992 534,263 2,960,468 1,541,923 631,302 222,035 49,557 27,648 6,449 2,910 475 6791,993 532,829 3,218,301 2,064,341 1,045,897 442,964 135,546 73,202 23,994 12,459 3,111 2,7191,994 802,548 4,485,852 2,717,809 1,453,372 664,828 229,411 124,384 44,503 23,932 7,979 5,4201,995 434,499 2,701,277 1,995,879 1,280,593 665,278 259,816 144,684 55,503 30,284 8,403 6,8861,996 284,596 2,081,852 1,854,836 1,466,609 870,358 403,111 237,378 101,414 56,838 20,823 13,8061,997 355,958 2,846,298 2,741,614 2,273,194 1,518,477 798,598 543,824 257,624 146,653 46,102 35,9301,998 186,691 1,709,738 1,983,000 2,015,324 1,591,306 961,818 674,205 349,447 203,445 75,899 43,3961,999 314,461 2,284,113 2,136,974 1,871,915 1,311,272 742,332 529,421 256,414 155,978 56,235 34,1882,000 165,964 1,638,777 2,123,191 2,408,398 2,125,559 1,341,381 956,664 456,514 287,609 98,578 63,9722,001 276,053 2,427,922 2,769,453 2,860,242 2,329,382 1,377,722 954,574 467,770 272,837 86,322 65,1292,002 308,909 2,612,711 2,774,499 2,534,875 1,864,685 1,020,896 688,551 310,252 187,810 59,341 42,672

154

Table G6. Selectivity patterns for the commercial and recreational fishery determined using the ‘un-tuned’ separable VPA.

Age 0 Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Age 7 Age 8 Age 9Age 10+

Com S(J) 0.036 0.383 0.606 0.809 1.000 1.000 1.000 1.000 1.000 1.000 1.000Rec S(J) 0.083 0.737 0.863 0.972 1.000 1.000 1.000 1.000 1.000 1.000 1.000Scrap S(J) 0.286 1.000 0.508 0.209 0.082 0.010 0.015 0.010 0.000 0.000 0.000Com/Scrap S(J) 0.110 0.614 0.717 0.855 1.000 1.000 1.000 1.000 1.000 1.000 1.000

Table G7. Selectivity patterns for the indices used in the revised model. Age 0 1 2 3 4 5 6 7 8 9 10+

NEFSC Trawl 0.79 1.00 0.40 0.34 0.11 0.12 0.00 0.00 0.00 0.00 0.00MRFSS 0.08 0.74 0.86 0.97 1.00 1.00 1.00 1.00 1.00 1.00 1.00

SEAMAP North 1.00 0.40 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00VIMS 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

155

Table G8. Recruitment deviations, fully selected fishing mortality rates, average fishing mortality rate (ages 1-10) and exploitation rate for Atlantic croaker in the mid-Atlantic. Estimates are for the base model (steepness=0.76; natural mortality=0.3).

Year

Commercial Recreational Scrap/Discard Average

F Exploitation

Recruitment deviations (Age 4) (Age 4) (Age 1) Ages 1-10 Rate

1973 0.10 0.04 0.13 0.17 0.101974 0.933 0.21 0.06 0.18 0.28 0.091975 1.433 0.35 0.08 0.06 0.27 0.091976 0.284 0.30 0.06 0.04 0.22 0.131977 -1.079 0.31 0.06 0.07 0.27 0.191978 -1.585 0.30 0.06 0.16 0.33 0.221979 -0.132 0.36 0.05 0.55 0.50 0.251980 -0.464 0.34 0.03 0.28 0.41 0.211981 -1.775 0.23 0.02 0.22 0.28 0.191982 -1.505 0.27 0.02 0.36 0.35 0.221983 1.395 0.19 0.02 0.18 0.23 0.071984 0.434 0.19 0.02 0.06 0.15 0.091985 0.683 0.14 0.01 0.05 0.12 0.061986 0.299 0.11 0.01 0.02 0.09 0.061987 0.278 0.07 0.01 0.05 0.08 0.051988 -0.914 0.06 0.02 0.06 0.09 0.061989 -0.403 0.04 0.01 0.08 0.06 0.051990 -0.620 0.03 0.01 0.07 0.05 0.041991 1.256 0.02 0.01 0.05 0.04 0.021992 0.143 0.02 0.01 0.02 0.03 0.021993 0.476 0.04 0.01 0.02 0.04 0.031994 0.340 0.04 0.01 0.03 0.05 0.031995 0.414 0.05 0.01 0.03 0.06 0.041996 -1.091 0.06 0.01 0.02 0.06 0.051997 0.070 0.09 0.02 0.03 0.10 0.061998 0.867 0.09 0.03 0.02 0.10 0.051999 0.799 0.09 0.02 0.02 0.09 0.052000 -0.224 0.10 0.03 0.01 0.09 0.062001 -0.504 0.10 0.03 0.01 0.11 0.082002 0.686 0.10 0.03 0.02 0.11 0.05

156

Table G9. Population estimates for the base mid-Atlantic model (steepness=0.76, natural mortality=0.3).

Population estimates

Numbers (millions)Biomass (metric

tons)

Year Age 0 TotalSpawning stock

biomass1973 50.31 145.10 13,1961974 297.37 392.17 13,1961975 489.01 748.66 21,2611976 181.15 677.65 38,3381977 52.53 478.51 45,1651978 32.46 311.33 39,8831979 135.57 307.21 30,0571980 93.08 255.17 22,3891981 23.44 166.07 18,5661982 29.28 125.13 15,3871983 500.84 570.42 11,7461984 175.21 565.68 27,9201985 287.40 663.85 37,8041986 207.52 664.10 49,0701987 210.79 671.51 57,3731988 65.43 534.06 63,8001989 109.91 476.13 62,2211990 88.40 422.25 60,0991991 572.80 871.25 57,0851992 187.78 818.23 73,2741993 270.26 862.02 82,9241994 238.21 856.89 90,0231995 258.43 868.30 94,1311996 57.35 672.92 96,6861997 183.93 654.18 89,6241998 403.39 853.69 81,5901999 375.06 974.35 84,4122000 135.18 817.37 91,0402001 103.10 663.92 88,7732002 338.55 786.87 80,328

157

Table G10. Summary statistics of average fishing mortality estimates from 2,500 simulations examining model sensitivity to steepness and natural mortality estimates. Percentile represents the estimate at the nth percentile

Year Percentile 100 90 75 50 25 10 0

1973 0.454 0.246 0.190 0.149 0.118 0.092 0.0011974 0.800 0.400 0.305 0.244 0.194 0.147 0.0021975 0.487 0.319 0.282 0.249 0.212 0.169 0.0021976 0.302 0.238 0.228 0.215 0.192 0.157 0.0021977 0.358 0.286 0.276 0.261 0.230 0.187 0.0021978 0.451 0.358 0.344 0.322 0.278 0.218 0.0021979 0.911 0.627 0.561 0.488 0.381 0.267 0.0021980 0.834 0.530 0.464 0.394 0.311 0.225 0.0021981 0.771 0.385 0.325 0.262 0.199 0.139 0.0011982 1.539 0.524 0.412 0.314 0.225 0.150 0.0011983 1.272 0.323 0.252 0.196 0.144 0.098 0.0011984 0.824 0.184 0.158 0.134 0.106 0.079 0.0011985 0.941 0.142 0.125 0.107 0.086 0.065 0.0011986 0.861 0.109 0.097 0.084 0.069 0.054 0.0011987 0.578 0.094 0.084 0.074 0.062 0.049 0.0011988 0.547 0.101 0.091 0.080 0.067 0.053 0.0011989 0.339 0.075 0.068 0.060 0.051 0.040 0.0011990 0.246 0.061 0.054 0.048 0.041 0.032 0.0001991 0.157 0.046 0.041 0.036 0.030 0.024 0.0001992 0.089 0.033 0.031 0.027 0.024 0.019 0.0001993 0.121 0.050 0.046 0.041 0.035 0.029 0.0001994 0.128 0.061 0.056 0.050 0.043 0.035 0.0011995 0.138 0.070 0.063 0.056 0.049 0.040 0.0011996 0.149 0.080 0.072 0.064 0.056 0.046 0.0011997 0.242 0.131 0.114 0.100 0.086 0.070 0.0011998 0.236 0.127 0.109 0.095 0.081 0.066 0.0011999 0.215 0.117 0.103 0.091 0.078 0.064 0.0012000 0.222 0.118 0.104 0.092 0.079 0.064 0.0012001 0.272 0.140 0.122 0.106 0.091 0.072 0.0012002 0.295 0.144 0.123 0.105 0.088 0.071 0.001

158

Table G11. Summary statistics of spawning stock biomass estimates (MT) from 2,500 simulations examining model sensitivity to steepness and natural mortality estimates. Percentile represents the estimate at the nth percentile

Year Percentile 100 90 75 50 25 10 01973 2,419,230 28,912 21,440 16,605 12,844 9,577 4,7831974 2,419,230 28,912 21,440 16,605 12,844 9,577 4,7831975 3,093,457 39,208 30,776 25,395 21,937 19,239 11,6601976 4,541,037 61,215 50,452 44,161 40,545 37,928 25,7681977 5,432,598 70,584 58,264 51,450 47,654 44,870 30,7921978 5,585,148 65,219 52,325 45,409 41,998 39,873 27,0621979 5,279,292 53,741 41,120 34,635 31,285 29,194 19,1401980 5,087,160 46,082 33,171 26,201 22,642 20,045 11,3281981 4,812,990 41,174 28,831 22,077 18,476 15,875 8,1251982 4,317,903 36,221 24,815 18,570 14,938 12,228 5,1251983 3,780,529 30,646 20,267 14,536 10,952 8,329 1,9761984 5,102,710 53,377 40,201 32,421 28,050 24,871 10,9441985 5,996,611 68,314 52,752 43,529 38,157 34,222 11,5191986 7,024,952 85,115 67,147 56,134 49,620 44,762 12,7351987 7,771,811 97,734 77,506 65,372 57,975 52,773 12,1261988 8,328,488 107,573 85,783 72,610 64,653 58,648 13,9581989 8,144,911 105,810 84,049 71,235 62,962 56,836 11,8161990 7,760,380 102,403 81,606 69,064 60,675 54,172 12,3231991 7,216,889 97,654 77,812 66,011 57,442 50,129 13,1851992 8,674,023 121,609 98,370 84,198 74,912 67,897 23,8141993 9,401,115 135,410 110,356 95,004 85,005 77,282 32,1941994 10,012,997 146,305 119,179 103,082 92,233 83,566 38,7271995 10,422,803 152,437 124,758 107,967 96,198 86,788 43,4281996 10,764,486 157,121 128,480 111,105 98,679 88,307 46,6851997 10,220,093 148,710 120,302 103,480 90,871 79,154 42,3151998 9,767,586 137,817 111,400 94,527 81,721 69,805 36,0951999 10,260,406 143,059 114,989 97,877 85,076 73,766 38,2442000 11,102,100 152,708 123,121 105,119 92,305 81,803 42,7962001 11,105,390 150,927 120,797 102,647 89,550 78,558 40,3092002 10,530,092 139,937 110,900 93,237 80,103 68,660 33,410

159

Table G12. Summary of sensitivity of reference point estimates in 2002 to varying weightings of likelihood components and alternate steepness and natural mortality estimates. For weighting sensitivity, the table summarizes the range of estimates from 3,500 runs. For the steepness and natural mortality sensitivity, the table summarizes the range of estimates from 2,500 runs. See text for details. Sensitivity to Statistic Percentile 100 90 75 50 25 10 0Weighting F2002: Fmsy 0.96 0.44 0.28 0.14 0.04 0.002 0.002Weighting SSB2002: SSBmsy 5.54 5.12 4.93 4.52 2.93 2.27 0.80Steep & MF2002: Fmsy 27.83 0.44 0.37 0.31 0.27 0.23 0.011Steep & MSSB2002: SSBmsy 8.13 2.97 2.78 2.61 2.41 2.16 1.50 Table G13. Sensitivity of reference point estimates and status of population in 2002 when estimates of Atlantic croaker bycatch from the shrimp fishery are included in the model. See text for details. Based on 1,00 runs. Percentile represents the proportion of run that had estimate equal to the estimate. e.g. For Fmsy , 90% of the runs had an Fmsy estimate <= 0.410. Percentile 100 90 75 50 25 10 0Fmsy 0.415 0.410 0.408 0.407 0.405 0.403 0.400F2002 0.176 0.129 0.118 0.105 0.094 0.082 0.057 SSBmsy (MT) 67,121 60,034 58,055 56,467 54,678 53,388 48,673SSB2002 (MT) 149,450 124,851 118,090 111,388 105,297 100,196 77,539 F2002:Fmsy 0.43 0.32 0.29 0.26 0.23 0.20 0.14SSB2002:SSBmsy 2.27 2.10 2.04 1.98 1.91 1.86 1.55

160

Figure G1. Standardized residuals for the commercial , recreational and scrap/discard landings for the mid-Atlantic base model (steepness=0.76 , natural mortality=0.3).

Commercial Landings

-4

-2

0

2

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000S

td. R

esid

ual

s

Recrational Landings

-2

0

2

4

6

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Std

. Res

idu

als

Scrap/Discard Landings

-4

-2

0

2

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Std

. Res

idu

als

161

Figure G2. Standardized residuals for the indices used in the mid-Atlantic base model (steepness=0.76, natural mortality=0.3).

NEFSC Trawl Survey

-4

-2

0

2

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000S

td. R

esid

ual

s

MRFSS Index

-4

-2

0

2

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Std

. Res

idu

als

SEAMAP Index

-4

-2

0

2

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Std

. Res

idu

als

VIMS Index

-4

-2

0

2

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Std

. Res

idu

als

162

Figure G3. Average fishing mortality rates (ages 1 –10) for Atlantic croaker in the mid-Atlantic. Figure G4. Spawning stock biomass (metric tons) and age 0 recruits (millions of fish) estimates from the base mid-Atlantic model ( steepness=0.76, natural mortality =0.3).

0.00

0.10

0.20

0.30

0.40

0.50

0.60

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Year

Ave

rag

e F

(1-

10)

0

20,000

40,000

60,000

80,000

100,000

120,000

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

SS

B (

met

ric

ton

s)

0

100

200

300

400

500

600

700

Ag

e 0

(nu

mb

ers

in m

illio

ns)

SSB Age 0

163

. Figure G5. (A). Average fishing mortality rates (ages 1-10+), from a simulation of the base model (steepness=0.76, natural mortality=0.3) using alternate weightings of the likelihood components (λ = 0 to 20, n=3,500). Circle represents median, box = inter-quartile range and lines 10-90th percentile. (B) Estimates for 2002, figure show the cumulative proportion of samples that had estimates at the 25th, 50 and 75th percentiles.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

73 75 77 79 81 83 85 87 89 91 93 95 97 99 1

Ave

rag

e F

(ag

es 1

-10+

)

0

0.2

0.4

0.6

0.8

1

0 0.1 0.2 0.3 0.4

F avg 02

Pro

po

rtio

n o

f va

lues

median 75th Perc 25th Perc

(B)

(A)

164

Figure G6. (A) Spawning stock biomass estimates (MT) from a simulation of the base model (steepness=0.76, natural mortality=0.3) using alternate weightings of the likelihood components (λ = 0 to 20, n=3,500). Circle represents median, box = inter-quartile range and lines 10-90th percentile. (B) Estimates for 2002, figure show the cumulative proportion of samples that had estimates at the 25th, 50 and 75th percentiles.

0

20,000

40,000

60,000

80,000

100,000

120,000

140,000

160,000

73 75 77 79 81 83 85 87 89 91 93 95 97 99 1

SS

B (

met

ric

ton

s)

0

0.2

0.4

0.6

0.8

1

0 50,000 100,000 150,000

SSB 02

Pro

po

rtio

n o

f va

lues

median 75th Perc 25th Perc

(A)

(B)

165

0.20.4

0.60.8

1.0

Steepness 0.206

0.301

0.397

Nat. Mort-81.52

-78.81

-76.09

-73.37

Tot Like

0.20.4

0.60.8

1.0

Steepness 0.210

0.305

0.400

Nat. Mort0

10

20

30

40

50

60

70number of runs

Figure G7. Sensitivity of model to varying steepness and natural mortality estimates: (A) Response surface of total likelihood estimates from 2,500 runs. Values towards low negative values (~ -73.37) represent the best fitting runs. (B) Sample size of steepness and natural mortality estimates evaluated. Steepness is binned in 0.1 intervals and natural mortality in 0.01 intervals.

(B)

(A)

166

Figure G8. (A) Average fishing mortality rates (ages 1-10+), from the 2,500 runs examining model sensitivity to steepness and natural mortality estimates. Circle represents median, box = inter-quartile range and lines 10-90th percentile. Steepness and natural mortality estimates were based on prior distributions from Myers et al (2002) for steepness and that developed by the TC for natural mortality. (B) Estimates for 2002, figure show the cumulative proportion of samples that had estimates at the 25th, 50 and 75th percentiles.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

73 75 77 79 81 83 85 87 89 91 93 95 97 99 01

F a

vger

age

(ag

es 1

-10+

)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

F avg 02

Pro

po

rtio

n o

f S

amp

les

median 75th Percentile 25th Percentile

(B)

(A)

167

Figure G9. (A) Spawning stock biomass estimates (MT) from the 2,500 runs examining model sensitivity to steepness and natural mortality estimates. Circle represents median, box = inter-quartile range and lines 10-90th percentile. Steepness and natural mortality estimates were based on prior distributions from Myers et al. (2002) for steepness and that developed by the TC for natural mortality. (B) Estimates for 2002, figures show the cumulative proportion of samples that had estimates at the 25th, 50 and 75th percentiles.

0

20

40

60

80

100

120

140

160

180

73 75 77 79 81 83 85 87 89 91 93 95 97 99 01

SS

B (

Th

ou

san

d M

etri

c T

on

s)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 50000 100000 150000 200000

SSB 02

Pro

po

rtio

n o

f S

amp

les

median 75th Percentile 25th Percentile

(B)

(A)

168

Figure G10. Stock –Recruit relationship fro Atlantic croaker using base model (steepness=0.76, natural mortality=0.3). Figure G11. Fishing mortality reference points relative to average fishing mortality rates across the time series for mid-Atlantic base model (steepness=0.76, natural mortality=0.3). Fmsy=0.39.

01

0099

9897

96

95

9493

92

91

90

8988

87

8685

84

83

82

81 80

79

78

7776

75

74

73

0

100

200

300

400

500

600

700

0 20000 40000 60000 80000 100000

Female Spawning Stock Biomass (mt)

Re

cru

its

(m

illio

ns

)

Estimated R S-R Model Replacement Line SSBmsy

0.00

0.10

0.20

0.30

0.40

0.50

0.60

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Ave

rage

F (ag

es 1

-10+

)

Threshold (Fmsy) Target 0.75Fmsy

169

Figure G12. Biomass reference points relative to SSB estimates for the mid-Atlantic base model (steepness=0.76, natural mortality=0.3). SSBmsy= 28,932 MT.

0

20,000

40,000

60,000

80,000

100,000

120,000

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

SS

B (

MT

)

Target (SSBmsy) Threshold 0.75 SSBmsy

170

Figure G13. Sensitivity of biological reference points (in 2002) to varying the weighting of the likelihood components for the base run (steepness=0.76, natural mortality=0.3). N= 3,500 runs. Figures show the cumulative proportion of samples that had estimates at the 25th, 50 and 75th percentiles.

00.10.20.30.40.50.60.70.80.9

1

0 0.2 0.4 0.6 0.8 1 1.2

F avg 02:F msy

Pro

po

rtio

n o

f v

alu

es

median 75th Perc 25th Perc

00.10.20.30.40.50.60.70.80.9

1

0 1 2 3 4 5 6

SSB 02:SSB msy

Pro

po

rtio

n o

f va

lues

median 75th Perc 25th Perc

(A)

(B)

171

Figure G14. Sensitivity of biological reference points (in 2002) to varying the steepness and natural mortality estimates used in model. N= 2,500 runs. Figures show the cumulative proportion of samples that had estimates at the 25th, 50 and 75th percentiles.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.2 0.4 0.6 0.8 1

F avg 02: F msy

Pro

po

rtio

n o

f S

amp

les

median 75th Percentile 25th Percentile

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.5 1 1.5 2 2.5 3 3.5 4

SSB 02: SSB msy

Pro

po

rtio

n o

f S

amp

les

median 75th Percentile 25th Percentile

(A)

(B)

172

Figure G15. Estimates of Atlantic croaker bycatch from the shrimp fishery used in the simulation analysis. The circle represents the median estimate of all runs, the box the inter-quartile range and the whiskers the range of estimates. The dark circle represents estimates from 1994.

0

2,000

4,000

6,000

8,000

10,000

12,000

73 75 77 79 81 83 85 87 89 91 93 95 97 99 1

AC

fro

m S

hri

mp

Bya

ctch

(M

T)

173

Figure G16. (A): Estimates of average fishing mortality rates (Ages 0-10+) and (B): spawning stock biomass (MT) for 2002 from sensitivity analysis where estimates of Atlantic croaker bycatch from the shrimp fishery are included in the model. Figures show the cumulative proportion of samples that had estimates at the 25th, 50, and 75th percentiles.

0

10

20

30

40

50

60

70

80

90

100

0 0.05 0.1 0.15 0.2

Avg F (0-10+) 2002

Cu

mu

lati

ve P

rop

. of

Sam

ple

s

75th Perc. 50th Perc. 25th Perc.

0

10

20

30

40

50

60

70

80

90

100

0 20000 40000 60000 80000 100000 120000 140000 160000

SSB 2002

Cu

mu

lati

ve P

rop

. of

Sam

ple

s

75th Perc. 50th Perc. 25th Perc.

(A)

(B)

174

Figure G17. (A): Estimates of Fmsy and (B) :SSBmsy from sensitivity analysis where estimates of Atlantic croaker bycatch from the shrimp fishery are included in the model. Figures show the cumulative proportion of samples that had estimates at the 25th, 50, and 75th percentiles.

0

10

20

30

40

50

60

70

80

90

100

0.38 0.39 0.4 0.41 0.42 0.43

F msy

Cu

mu

lati

ve P

rop

. of

Sam

ple

s

75th Perc. 50th Perc. 25th Perc.

0

10

20

30

40

50

60

70

80

90

100

30,000 35,000 40,000 45,000 50,000 55,000 60,000 65,000 70,000

SSB msy

Cu

mu

lati

ve P

rop

. of

Sam

ple

s

75th Perc. 50th Perc. 25th Perc.

(B)

(A)

175

Figure G18. Estimates of F2002: Fmsy and SSB2002: SSBmsy ratios from sensitivity analysis where estimates of Atlantic croaker bycatch from the shrimp fishery are included in the model. Figures show the cumulative proportion of samples that had estimates at the 25th, 50, and 75th percentile.

0

10

20

30

40

50

60

70

80

90

100

0 0.1 0.2 0.3 0.4 0.5

F 2002: F msy

Cu

mu

lati

ve P

rop

. of

Sam

ple

s

75th Perc. 50th Perc. 25th Perc.

0

10

20

30

40

50

60

70

80

90

100

1 1.2 1.4 1.6 1.8 2 2.2 2.4

SSB 2002:SSB msy

Cu

mu

lati

ve P

rop

. of

Sam

ple

s

75th Perc. 50th Perc. 25th Perc.

(B)

(A)

176

Appendix H: Status of stock identification of Atlantic croaker along the east coast on the U.S. A realistic population assessment has as its basis the correct identity of the species. Historically, there have been problems in the specific identification of exploited fishes. For example, redfish of the genus Sebastes occur in the deeper waters off the New England states and Canada, across the North Atlantic to Norway and the Barents Sea. Initially redfish were assigned to the species Sebastes marinus. Subsequent taxonomic studies found that there were two species: S. mentella and S. fasciatus. Because of this systematic problem, published and unpublished studies of “S. marinus” are suspect and may apply to either species. It is even more problematical in that they co-occur and also have been shown to hybridize (see Marcogliese et al., 2003). For a valid species with a broad distributional range, the next problem is to determine if, within its range, that species is a unit stock. This term has a variety of definitions. Ricker (1975) called it “a part of a fish population which is under consideration from the point of view of actual or potential utilization”. This definition is quite vague and has not been considered as the most appropriate. One of the most complete and understandable explanations of the term is that of Sparre and Venema (1992). From their perspective, a “stock is a sub-set of one species having the same growth and mortality parameters, and inhabiting a particular geographical area”. Stocks show little mixing with the adjacent group, and since they possess the same growth and mortality parameters over their distribution, they can undergo modeling of their population. According to Hilborn and Walters (1992) “a unit stock is an arbitrary collections of populations of fish that is large enough to be essentially self-reproducing (abundance changes are not dominated by immigration or emigration), with members of the collection showing similar patterns of growth, migration and dispersal”. Methods to determine the unity of a stock of fishes include tag-recapture, meristics and morphometrics, parasite studies, scale and/or otolith pattern analysis, and genetic analysis including blood protein electrophoresis and DNA analysis. Recently, advances in the chemical composition (elemental composition) of otoliths have provided a method to determine the estuarine origin as well as a check for fidelity of the early life stages to a given estuary. In addition, investigators have compared life history traits such as growth, maturity schedules, and physiological tolerance ranges to determine if these vary between areas within the species geographic range. Tag-recapture There are numerous articles and reviews of the problems associated with and the use of tag-recapture experiments on fishes. Fish are caught with the least ‘offensive’ gear, handled as delicately as possible and marked with tags that are of the proper size and construction. By the latter, we mean that the tag should be made of materials that last. I have marked fish with a type of tag (T-bar) in the proper place (locked in the pterygiophores of the dorsal fin) and had them fail in a very short time. The plastic shaft with the “T” separated from the part of the tag that had the sequential number as

177

well as the information for reporting the tag. To investigate the stock structure (or lack of structure) of Atlantic croaker along the US east coast, the tagging would occur to the south and north of the North Carolina coast to see if there is exchange between the two areas. If a sufficient number of fish were tagged and there was a reasonable reporting rate of the captured tagged fish, one could determine if there were northern and southern ‘stocks’ of Atlantic croaker. The Georgia Department of Natural Resources conducted a study of the inshore fish species of importance to recreational anglers (Music and Pafford 1984). The Atlantic croaker was included in the species list. Fishes were collected with a variety of gear, studied for biological characteristics and tagged and released in an attempt to get information on movements. They tagged 3,456 Atlantic croaker from April 3, 1979 through June 28, 1982. Only 2.5% of the tags were returned (n = 87). Returns were made by recreational anglers (n = 50), commercial fishermen (n = 13) as well as recaptures by project personnel (n = 24). Time at large (period between marking and return) ranged from 2 to 416 days with an average of 63 days. The longest distance traveled from the tagging to the recapture location was 179 km with an average of 10.9 km. Over half of the returns (50.6%) were from fish caught in the same general location as where marked and released. Most of the recaptures came from either the creek systems or the Georgia Sounds. Of the three fish that moved over 100 km, two moved south to the St. Johns River in Florida and one moved north to Cane Island South Carolina. The remaining recaptures were less than 50 km from the release location. A longterm study of the biota of the Cape Fear River, North Carolina estuarine system was conducted by Swartz et al. (1979). One aspect of this work was to examine movements of fishes within and between areas by tagging. Over the study period (1973 – 1977), 28,231 Atlantic croaker were tagged and released. Approximately 2% of the tags (n = 563) were returned to the investigators. Only 14 of the returns showed significant movement along the coast from 75 to ~300 km. Six fish were taken between the Cape Fear system and Cape Lookout, NC; seven fish between Cape Lookout and Ocracoke Island, NC; a single fish was caught between Ocracoke Island and Cape Hatteras. No fishes were caught at significant distances to the south of Cape Fear. Thus, some fishes tagged in the areas around Cape Fear moved as far north as the area around Cape Hatteras indicating that there was the potential for movement into the Mid-Atlantic region. Although there may be other studies conducted by various states and agencies, we could find no other results for the east coast of the United States. Morphometrics and meristics In the recent literature, a study by Ross (1988) suggested that the growth rates of Atlantic croaker from the coastal areas of North Carolina showed differences. The more northerly group (northern North Carolina to Chesapeake Bay) lived longer and had a different growth curve than the more southerly group (south of Pamlico Sound,

178

North Carolina). He suggested that the two groups over-lapped off North Carolina. It is difficult to determine from his paper whether he considered these “groups” as separate stocks. He did suggest, however, that the northern “group” had differing life history traits such as a longer life span, a later age at sexual maturity than the southern group. The latter was characterized by faster growth, early maturation, a shorter life span, and smaller size. Two comments concerning this work. First, ages of the Atlantic croaker used in his study were determined from the analysis of scales. Second, the maturity schedules based on scale ages are suspect and those reports from the literature need to be re-evaluated. The last sentence in his paper is “Population dynamics and resulting fishery management in North Carolina may be confounded by a mixing of Atlantic croaker stocks until adequate separation techniques are developed.” I argue that he has not provided a sufficiently detailed analysis of the species throughout its distributional range along the east coast of the United States to define two “groups” or stocks. Genetic analysis Recent advances in analytical techniques to determine the composition of the genetic materials of species and individuals within species has led to a blossoming of studies attempting to determine the stock structure of marine fishes. Lankford et al. (1999) applied the analysis of mitochondrial DNA to investigate the possible stock structure of Atlantic croaker in U.S. waters. Their main finding was “MtDNA analysis provided no evidence that M. undulatus is subdivided by Cape Hatteras into discrete genetic stocks. Frequency and distance-based analyses both suggested a single, panmictic population of Atlantic croaker on the U.S. Atlantic coast.” Lankford et al. (1999) did find differences between fishes from the eastern Atlantic coast and those from the Gulf of Mexico. As is the case of most genetic analyses, the paper concluded with the following statement: “Mark-recapture studies designed to quantify the level of adult migration across Cape Hatteras, combined with otolith-microchemical analyses to examine larval dispersal patterns, could provide valuable information on the level of mixing between the MAB and South Atlantic Bight (SAB) areas and clarify the extent to which M. undulatus in these regions constitute self-recruiting groups.” The use of modern, genetic analyses has failed to prove that the Atlantic croaker along the east coast of the United States forms “group” or stocks. Lankford et al. (1999) also state that “Because low levels of gene flow may produce mtDNA homogeneity between otherwise self-recruiting stocks, mtDNA is incapable of distinguishing between low (1%) and moderate (50%) amounts of mixing.” No doubt, some new more sensitive technique will come into vogue and all these studies will be repeated.

179

Elemental composition of otoliths Thorrold et al. (1997) applied laser ablation inductively coupled plasma mass spectrometry (LA – ICPMS) to the sagittae of wild caught Atlantic croaker juveniles. These fish originated from the Neuse River, NC and the Elizabeth River, VA. The technique basically vaporizes a very small area of the otolith and passes the resulting gas through a mass spectrometer to determine the concentrations and presence of specific elements. The investigator is then able to compare the elemental composition between areas or across different regions of the otolith section to determine ontogenetic changes in the elements and their concentrations. Since the elements are absorbed from seawater and incorporated into the structure of the otolith as it grows, the elemental composition of the otolith reflects the characteristics of the water body where the fish was at that time. Thorrold et al. (1997) tested the sagittae of fishes from the above areas and found no differences in the chemical make up of the core area of the sagittae. The Atlantic croaker spawns off-shore on shelf and near-shore waters during the late summer and fall off the Middle Atlantic states and in off-shore waters primarily in late-fall and winter in the South Atlantic Bight. Their test fishes were caught in March and April 1994 and ranged between 20 and 45-mm standard length. The analyses showed that the chemical signatures of the central regions of the otoliths from the two areas could not be separated. The centers are deposited in the first days of the larval fishes’ life and indicate the water mass within which the spawning event took place. The signatures suggested that they all originated from the same area. The authors stated “We were, however, unable to reject the hypothesis that Atlantic croaker larvae from north and south of Cape Hatteras originated from different spawning sites. This may indicate that the larvae were spawned in close geographical proximity, and strengthens arguments that Atlantic croaker in the MAB and SAB represent a single spawning stock.” Thus, the use of highly sophisticated analytical techniques were able to define “groups” of Atlantic croaker along the east coast of the U.S. One question jumps to mind from the paper. The Neuse River can not be referred to as ‘north of Cape Hatteras” since the North Carolina Sounds are all connected and the juveniles may have gained access to the sounds from the inlets north of the Cape, i.e., Oregon Inlet. Secondly, even if the Neuse River is considered to be “south of Cape Hatteras”, it would have been very helpful to substitute fishes from the Cape Fear River, North Carolina or some estuarine system further to the south. The stock definition issue is clouded between Atlantic croaker found south of Cape Fear, NC to near Cape Canaveral, FL and those of the Middle Atlantic such as Chesapeake Bay, VA and Delaware Bay. Analysis of the parasitic fauna The analysis of the parasitic fauna on fishes has been used for quite some time to study populations and movements of marine fishes (see MacKenzie 1983 for a review). When a fish becomes infected with a parasite, it contacts the parasite’s infective phase. This may occur when a fish swallows food items that act as vectors, the parasite

180

attaches to the body of the host, or when the fish becomes inoculated by a tissue feeder such as a leech (Lester 1990). A particular parasite generally has a restrictive distribution over the host’s range because there are temporal and spatial limitations when a host is in proximity to the infective stage. As Lester (1993) so aptly points out “As fish move into the exchange points, they become infected, and as they move out, they carry a legacy of their occupancy of the points.”. The use of parasites as indicators of population structure and fish movement is not without limitations. Sinderman (1957) presented the following as contributing to these limitations: (1) the variability of parasitic infection with season and location; (2) does a parasitic infection cause differential mortality so that the resulting sampled group of fish is not representative to the real population; (3) the distribution of the parasite within a host may reflect variations in the distributions or abundance of intermediate or alternate hosts; (4) environmental factors such as temperature, salinity, etc.; and (5) longterm fluctuations may not be apparent in short term studies. To be used in movement and population studies the following factors need to be addressed: (1) host specificity – does the parasite infect the target species and how specific is this parasite to the host; (2) what is the geographic distribution of the host as well as the parasite; (3) what is the sex of the host; (4) what is the size of the host; (5) what is the season of examination of the potential host for the presence of the parasite; and (6) what is the location of infection on the host. In a comparison of the parasitic fauna infecting Atlantic croaker from various latitudes along the east coast of the US, Thoney (1993) reports some differences in the abundance of a suite of parasites in Atlantic croaker. The specimens were taken during the spring and fall groundfish surveys conducted along the east coast of the US by the National Marine Fisheries Service, Woods Hole. A weakness in this specific work resulted from the design of the research. Fishes were taken in two seasons, however, during the cooler months (i.e., spring), Atlantic croaker are in the warmer waters of the southern part of the survey’s range. As water temperatures warm, this species moves north and inshore. Essentially, this study compared the parasites of the same “groups” of fish and the resulting differences may have been seasonal, size, or age related and hence could not form a solid foundation for stock determination. The distributional range of the samples was restricted to the northern and central part of the range along the east coast. Researchers at the College of Charleston and the South Carolina Department of Natural Resources (SCDNR) are investigating whether the parasitic fauna of Atlantic croaker can be used to group the Atlantic coast population into discrete units. Fishes are collected during the groundfish survey of the Northeast Fisheries Science Center of the NMFS. These are frozen and shipped to the laboratory in Charleston, SC. Fishes are measured, weighed, sexed, assigned a reproductive condition, and aged by thin sections of sagittal otoliths. A variety of tissues and organ systems are examined for infection. The parasites are identified to the lowest possible taxon, counted and the within host distribution is documented.

181

Preliminary results are promising, however, the study has been on-going for one year and has two additional years to completion. Below is the general summary of the parasite information to date. The data will be presented at a scientific meeting. Remember, these are preliminary and fishes from other seasons need to be analyzed. Abstract - presentation for SSP - April 04 In order to identify potential stock populations of the Atlantic croaker, Micropogonias undulatus, on the eastern coast of the United States, SCDNR initiated the study of the macroparasite fauna of this fish species to determine if some of its parasites could be used as biological tags. This is a three year project (2002-2005) and results presented herein are restricted to findings generated from the dissection of 111 Atlantic croakers collected from the New Jersey coast through Cape Canaveral FL in the Fall of 2002. Of all the macroparasite species collected, 2 acanthocephalans (Pomphorhynchus rocci and Serrasentis sagittifer), 1 nematode (Spirocamallanus cricotus), 1 cestode (Scolex polymorphus unilocularis), 1 copepod (Lernaeenicus radiatus), 1 digenean (Diplomonorchis leistomi), and 1 gastric monogenean (species yet to be identified and described) showed differences in occurrence north and south of Cape Hatteras, N.C., and are thus considered to be good candidates to act as natural tags marking potential Atlantic croaker stocks. Funded by a MARFIN grant NA17FF2885. Comparison of life history traits Several researchers have suggested that there are two stocks of Atlantic croaker along the east coast of the US with Cape Hatteras forming the breaking point between these stocks. They have indicated that the fish north of Hatteras have a longer life span, later age and size at sexual maturity, and differing mortality schedules and growth rates. The problem was that in these studies, the basis for comparison was suspect. Maturity schedules, mortality rates, longevity, and growth all require the proper determination of age. The initial work was based on the analysis of scales (see Ross 1988). Subsequent work has used otoliths, however, the location and description of the first annulus was difficult to reproduce and requires standardization (see Barbieri et al. 1994). Also, the latter work was based on specimens from only the Chesapeake Bay area and renders the discussion and the interpretation of the findings comparing other geographical locations as questionable. Lankford and Targett (2001) conducted a series of tests on age-0 Atlantic croaker to determine their environmental tolerances and make comparisons along a latitudinal gradient along the east coast of the US. The initial work established the rates of survival at different temperatures. It ranged from 0% at 1oC to 99.3% at 7oC. The survival rate dramatically increased between 3 oC (1.3%) and 5oC (86.8%). They also found that size had an impact on survival with smaller individuals being able to survive longer than larger individuals. Also, at higher salinities, survival increased. The next

182

series of experiments compared these findings from age-0 Atlantic croakers collected in different Atlantic Coast estuaries (Lankford and Targett 2001). The sites were Delaware Bay, DE, Cape Fear River, NC, and Indian River Lagoon, FL. Growth capacity, feeding rate, growth efficiency, and cold tolerance were similar across geographic locations. This provided supporting evidence of a single genetic stock of Atlantic croakers along the US east coast. Summary One stock, two stock, three stock, four stock The growth data as well as other life history comparisons on a latitudinal gradient along the east coast of the U.S. are suspect because of age problems (scales against otolith sections; differing interpretations of the first check mark on the sections). The tagging data from in the lower portion of the SAB suggests that although there is movement between states, no long-distance movements north of Cape Romain, SC were noted. On the other hand, the tagging information from the Cape Fear River, NC study shows movement from Cape Fear to the area around Cape Hatteras, NC indicating that the movement of fishes from the SAB may occur, but its significance is not known. The use of genetics has failed to show any differences between fishes in different areas along the east coast. Only a small amount of interchange obscures population differences. So far, the Atlantic croaker found along the east coast of North America may form two separate stocks; those of the Gulf of Mexico differ genetically from those along the east coast of the US and analytical techniques are not sophisticated enough to determine if there indeed are separate groups along the east coast. The parasite data are incomplete and require more locations, sizes and sexes to fill out the various sampling categories. The temperature tolerances, i.e., survival at low temperatures, are similar along the east coast for young-of-year (YOY). Growth parameters for YOY are also consistent throughout the region. Future studies of various traits should examine fishes from the limits of their distributional range along the east coast of the US, and then examine those between the extremes to determine if there are indeed two stocks of Atlantic croaker. Literature Cited BARBIERI, L.R., M.E. CHITTENDEN,JR., AND C.M. JONES.

1994. Age, growth and mortality of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay region with a discussion of apparent geographic changes in population dynamics. Fish. Bull. 92:1-12.

HILBORN, R. AND C.J. WALTERS. 1992. Quantitative Fisheries Stock Assessment. Chapman and Hall, NY,

NY. 570p.

183

LANKFORD, T.E., JR., AND T.E. TARGETT. 2001. Low-temperature tolerance of age-0 Atlantic croakers: recruitment

implications for U.S. Mid-Atlantic estuaries. Trans. Amer. Fish. Soc. 130:236-249.

LANKFORD, T.E., JR., AND T.E. TARGETT. 2001. physiological performance of young-of-year Atlantic croakers from

different Atlantic coast estuaries: implications for stock structure. Trans. Amer. Fish. Soc. 130: 367-375.

LANKFORD,T.E.,Jr, T.E. TARGETT AND P.M.GAFFNEY. 1999. Mitochondrial DNA analysis of population structure in the Atlantic

croaker, Micropogonias undulatus (Perciformes: Sciaenidae). Fish. Bull. 97: 884-890.

LESTER, J.G. 1990. Reappraisal of the use of fish parasites for fish stock identification.

Aust. J. Mar. Freshwater Res. 41:855-864. MacKENZIE,K. 1983. Parasites as biological tags in fish population studies. Advances in

Appl. Biol. 7:251-331. MARCOGLIESE,D.J., E. ALBERT, P. GAGNON, J. SEVIGNY. 2003. Use of parasites in stock identification of the deepwater redfish

(Sebastes mentella) in the Northwest Atlantic. Fish. Bull. 101:183-188. MUSIC, J.L.,JR AND J.M. PAFFORD. 1984. Population dynamics and life history aspects of major marine

sportfishes in Georgia’s coastal waters. GA DNR, Coastal Resources Dic., Contri. Ser. No. 38, 382p.

RICKER, W.E. 1975. Computation and interpretation of biological statistics of fish

populations. Fish. Res. Bd. Canada Bull. 191, 382 p.

ROSS,S.W. 1988. Age, growth, and mortality of Atlantic croaker in North Carolina, with

comments on population dynamics. Trans. Amer. Fish. Soc. 117: 461-473. SINDERMAN,C.J. 1983. parasites as natural tags for marine fish: a review. Northwest

Atlantic Fish. Org. Sci. Council Ser. 6: 63-71. SPARRE, P. AND S.C. VENEMA. 1992. Introduction to tropical fish stock assessment. Part 1. Manual. FAO

184

Rome, 376p. SWARTZ, F.J., L. DAVIDSON, C. SIMPSON, M. McADAMS,K. SANDOY, J. DUNCAN AND D. MASON.

1979. An ecological study of fishes and invertebrate macrofauna utilizing the Cape Fear River, Carolina Beach Inlet, and adjacent Atlantic Ocean. Inst. Mar. Sci. Univ. North Carolina. Final Report to Carolina Power and Light Company, Raleigh, N.C.,571 p.

THONEY, D.A. 1993. community ecology of the parasites of adult spot, Leiostomus

xanthurus, and Atlantic croaker, Miscropogonias undulatus (Sciaenidae) in the Cape Hatteras region. J. Fish. Biol. 43: 781-804.

THORROLD,S.R.,C.M.JONES, AND S.E. CAMPANA. 1997. response of otolith microchemistry to environmental variations

experienced by larval and juvenile Atlantic croaker (Micropogonias undulatus). Limnol. Oceanog. 42: 102-111.

Atlantic States Marine Fisheries Commission

Atlantic Croaker 2003 Stock Assessment Report

DRAFT

October 2003

Working towards healthy, self-sustaining populations for all Atlantic coast fish species

or successful restoration well in progress by the year 2015

2

Table of Contents Terms of Reference ....................................................................................................................... 10 1.0 Introduction ............................................................................................................................. 10 2.0 Life History ............................................................................................................................. 11

General Information .................................................................................................................. 11 2.1 Age ...................................................................................................................................... 12 2.2 Growth ................................................................................................................................ 13 2.3 Reproduction ....................................................................................................................... 13

2.3.1 Sex ratio ....................................................................................................................... 14 2.3.2 Size and Age at Maturity ............................................................................................. 14

2.4 Stock definitions ................................................................................................................. 14 3.0 Fishery Description ................................................................................................................. 15

3.1 Brief overview of Fisheries ................................................................................................. 15 3.1.1 Commercial Fishery ......................................................................................................... 15 3.1.2 Recreational Fishery ........................................................................................................ 16

3.2 Regulations and Management History .................................................................................... 16 4.0 Habitat Description ................................................................................................................. 17 5.0 Data Sources ..................................................................................................................... 18

5.1 Commercial ......................................................................................................................... 18 5.1.1 Data Collection Methods ............................................................................................. 18

5.1.1.1 Survey Methods .................................................................................................... 18 5.1.1.2 Sampling intensity ................................................................................................ 19 5.1.1.3 Biases .................................................................................................................... 19 5.1.1.4 Aging methods ...................................................................................................... 19

5.1.2 Commercial landings ................................................................................................... 20 5.1.3 Commercial discards and bycatch ............................................................................... 21 5.1.4 Commercial catch rates ................................................................................................ 21 5.1.5 Commercial catch at age .............................................................................................. 21

5.2. Recreational ....................................................................................................................... 21 5.2.1 Data Collection Methods ............................................................................................. 22

5.2.1.1 Survey Methods .................................................................................................... 22 5.2.1.2 Sampling Intensity ................................................................................................ 22 5.2.1.3 Biases .................................................................................................................... 22 5.2.1.4 Biological Sampling .............................................................................................. 23 5.2.1.5 Aging Methods...................................................................................................... 23 5.2.1.6 Development of Estimates .................................................................................... 23

5.2.2 Recreational Landings ................................................................................................. 24 5.2.3 Recreational Discards .................................................................................................. 25 5.2.4 Recreational Catch Rates ............................................................................................. 26 5.2.5 Recreational Catch-at-Age ........................................................................................... 28

5.3 Fishery- Independent Survey data ..................................................................................... 28 5.3.1 SEAMAP ..................................................................................................................... 28

5.3.1.1 Sampling Intensity ................................................................................................ 28 5.3.1.2 Biological Sampling .............................................................................................. 29 5.3.1.3 Aging Methods...................................................................................................... 30

3

5.3.1.4 Development of Estimates .................................................................................... 30 5.3.2 NMFS Northeast Trawl Survey ................................................................................... 30

5.3.2.1 Sampling Intensity ................................................................................................ 30 5.3.2.2 Biases .................................................................................................................... 30 5.3.2.3 Aging Methods...................................................................................................... 31 5.3.2.4 Development of Estimates .................................................................................... 31

5.3.3 Virginia Institute of Marine Sciences (VIMS) Trawl Survey. ..................................... 31 5.3.3.1 Sampling Intensity ................................................................................................ 31 5.3.3.2 Biases .................................................................................................................... 31 5.3.3.3 Biological Sampling .............................................................................................. 31 5.3.3.4 Aging Methods...................................................................................................... 31 5.3.3.5 Development of Estimates .................................................................................... 31

5.3.4 Length/Weight/ Catch-at-Age ...................................................................................... 32 5.3.5 Abundance Indices ....................................................................................................... 32 5.3.6 Biomass Indices ........................................................................................................... 32 5.3.7 Natural Mortality Estimates ......................................................................................... 32

6.0 Methods................................................................................................................................... 32 6.1 Model(s) .............................................................................................................................. 32

6.1.1. Surplus Production Model .......................................................................................... 33 6.1.2. Age Structured Production Model. ............................................................................. 33

6.2 Model Calibration ............................................................................................................... 36 6.2.1 Tuning Indices ............................................................................................................. 38

6.2.2. Input Parameters and Specifications ............................................................................... 38 Selectivity ................................................................................................................................. 39 Biological characteristics .......................................................................................................... 40 Natural mortality and steepness ................................................................................................ 40 Weighting of the likelihood components .................................................................................. 41 Bounds for parameters estimated by the model ........................................................................ 41

7.0 Outputs/Results ....................................................................................................................... 41 7.1 Goodness of Fit of Model Used .............................................................................................. 42 7.2 Parameter Estimates ................................................................................................................ 42

7.2.1 Exploitation Rates (should include both F and u) ........................................................ 42 7.2.2 Abundance Estimates ................................................................................................... 43 7.2.3 Precision of Parameter Estimates................................................................................. 44

7.3 Projection Estimates............................................................................................................ 44 7.4 Sensitivity Analyses ............................................................................................................ 44 7.5 Retrospective Analyses ....................................................................................................... 44

8.0 Biological Reference Points .................................................................................................... 45 8.1 Overfishing Definition ........................................................................................................ 45 8.2 Stock Recruitment Analysis ................................................................................................ 46 8.3 Yield and SSB per Recruit .................................................................................................. 46 8.4 Stock Production Model ..................................................................................................... 46

9.0 Recommendations and Findings ............................................................................................. 46 9.1 Evaluation of current status based on biological reference points ...................................... 47 9.2 Research Recommendations ............................................................................................... 47

10.0 Literature Cited ..................................................................................................................... 48

4

11.0 Tables .................................................................................................................................... 51 12.0 Figures................................................................................................................................... 99 Appendix A. ................................................................................................................................ 134 Comparison of Estimates using the Excel and AD model Builder age structured production model when similarly configured (base Mid-Atlantic model) .................................................... 134 Appendix B. Ad model builder template file used in analysesDATA_SECTION ..................... 135

5

List of Tables

Table 122.1.1. Summary of available age data for Atlantic croaker ( only samples based on otolith readings were considered). ..................................................................................... 51

Table 32.1.2. Summary of age structure of Atlantic croaker obtained from the available data sets ........................................................................................................................................ 52

Table 42.2.1. Summary of Von Bertalanffy Growth parameters examined for use in this assessment. Only studies base on age determination made using sectioned otoliths were considered. ........................................................................................................................... 53

Table 53.1.1.1: Commercial Landings of Atlantic Croaker in Pounds by Atlantic Coastal States, 1950-2001 ................................................................................................................. 54

Table 63.1.1.2 Commercial value of landings by state of Atlantic croaker ............................. 56

Table 73.2: Summary of current regulations for Atlantic croaker .......................................... 59

Table 85.1.2.1 Percent landings by gear for Atlantic coast commercial Atlantic croaker harvest .................................................................................................................................. 60

Table 95.1.2.2 Percent Commercial landings of Atlantic croaker, by state, 1950 - 2001 for Atlantic coast states ............................................................................................................. 62

Table 105.2.1.2 . Number of Intercept Trips in which Atlantic croaker could have been potentially caught but were not caught (zero trips), the number of intercepts where Atlantic croaker were caught (Positive Trips) and the number of Atlantic croaker measured by Region. ........................................................................................................... 64

Table 115.2.1.6. Size categories used to determine recreational discard weights ..................... 65

Table 125.2.2.1. Recreational Landings (Type A+B1 in numbers) of Atlantic croaker ........... 69

Table 135.2.2.2. Recreational Landings (Type A+B1 in pounds) of Atlantic croaker ............. 71

Table 145.2.2.4. Percentage of recreational landings by area and mode fished and total landings (numbers) ............................................................................................................. 73

Table 155.2.2.5. Estimated total recreational effort and targeted croaker trips by region. .... 74

Table 16175.2.2.6. Size distribution of Atlantic croaker weighted by landings (numbers) for the Mid Atlantic region of the fishery (North Carolina and north). Size class in 10 mm intervals is the lower bound. .............................................................................................. 75

Table 185.2.2.7. Size distribution of Atlantic croaker weighted by landings (numbers) for the South Atlantic region of the fishery (South Carolina and south). Size class in 10 mm intervals is the lower bound. .............................................................................................. 76

Table 195.2.3.1. Numbers of Atlantic croaker released alive by recreational fishermen (Type B2). ........................................................................................................................................ 77

Table20 5.2.3.2. Estimates number and weight of recreational discards. Discard weight (pounds) were estimated using seven methods described in the text. ............................ 78

6

Table 215.2.4.1. Species used to identify a potential Atlantic croaker intercept by state. The species most likely to be associated with an Atlantic croaker target trip were determined using a jaccard type index. ............................................................................ 79

Table22 5.2.4.2. Summary statistics for the negative binomial generalized linear model and log transformed general linear models used to estimate recreational catch rates ........ 81

Table 235.2.4.3. Estimates of recreational catch rates and 95% confidence intervals for Atlantic croaker in the Mid Atlantic (North Carolina and North) and South Atlantic regions (South Carolina and south) ................................................................................... 82

Table24 5.3.2.4.1 Estimates of catch per tow in numbers and weight for the NMFS trawl survey using the delta-log normal GLM. ......................................................................... 83

Table 255.3.3.5.1 Spring Atlantic Croaker (Recruit) Indices. Estimates of catch per tow in numbers for the VIMS trawls survey (spring). Data courtesy of VIMS. ...................... 84

Table 265.3.7.1. Mortality estimates for Atlantic croaker based on different studies and methods ................................................................................................................................ 85

Table 27 6.2.1.1. Summary Table of Available fishery Independent and dependent indices .. 87

Table 28 6.2.2.1. Selectivity estimates used in the base age structured production model ...... 88

Table29 6.2.2.2. Parameter bounds used in the AD model Builder version of the age structured production model . ........................................................................................... 88

Table30 7.1.1. Standardized residuals for the commercial and recreational landings for the Mid Atlantic and South Atlantic base models .................................................................. 89

Table31 7.1.2. Standardized residuals for the indices used in the base models for the Mid-Atlantic and South Atlantic models ................................................................................... 90

Table 327.2.1.1. Fully recruited fishing mortality estimate for Atlantic croaker from the base Mid-Atlantic and South Atlantic models. ......................................................................... 91

Table33 7.2.1.2. Exploitation rates for Atlantic croaker from the base mid-Atlantic and South-Atlantic models ........................................................................................................ 92

Table34 7.2.2.1. Population estimates for Atlantic croaker from the base mid-Atlantic and South-Atlantic models. ....................................................................................................... 93

Table 357.4.1. Summary of 1000 Monte-Carlo Trials to evaluate uncertainty surrounding the mid-Atlantic model. Estimates from the base mid-Atlantic model are included for comparative purposes. ........................................................................................................ 94

Table36 8.1.1. Biological Reference Points for Mid-Atlantic region. ......................................... 98

7

List of Figures

Figure 15.1.2.1 Atlantic coastal commercial landings of Atlantic croaker (metric tons), 1950-2001. ...................................................................................................................................... 99

Figure 25.2.2.1. Recreational landings of Atlantic croaker (numbers) by region. ................. 99

Figure 35.2.2.2. Recreational landings of Atlantic croaker (pounds) by region. .................. 100

Figure 45.2.2.3. Recreational landings by area fished and total landings (numbers) ....... 100

Figure 55.2.2.4. Recreational landings by mode fished and total landings (numbers) ....... 101

Figure 65.2.2.5. Proportion of Atlantic croaker landings by Wave and year. ...................... 101

Figure7 5.2.2.6. Estimated number of total recreational trips and trips targeting Atlantic croaker by region. ............................................................................................................. 102

Figure 85.2.2.7. Size distribution of Atlantic croaker for the northern region (North Carolina and North). ........................................................................................................ 102

Figure 95.2.2.8. Size distribution of Atlantic croaker for the southern region (South Carolina and South). ......................................................................................................................... 103

Figure10 5.2.3.1. Ratio of Atlantic croaker released by anglers to those landed. ................... 103

Figure11 5.2.4.1. Recreational catch rates and 95% confidence intervals for Atlantic croaker in the Mid Atlantic region (North Carolina and North) using a negative binomial generalized linear model and log transformed general linear model. .......................... 104

Figure 125.2.4.2. Recreational catch rates and 95% confidence intervals for Atlantic croaker in the South Atlantic region (South Carolina and South) using a negative binomial generalized linear model and log transformed general linear model. .......................... 104

Figure13 6.2.1.Normalized estimates for the two major fishery independent indices . NMFS=NEFFC trawl survey, SEAMAP= SEAMAP trawl survey. ............................. 105

Figure 146.2.2 Normalized fishery independent CPUE estimates (by Strata). ....................... 105

Figure 156.2.3. Posterior probability distributions for steepness at varying level of natural mortality used in the core models for the Mid Atlantic region (North). Prior probability distribution based on Myers et al . (2002) is also included. ...................... 106

Figure16 6.2.4. Posterior probability distributions for steepness at varying level of natural mortality used in the preliminary models for the South Atlantic region. Prior probability distribution based on Myers et al . (2002) is also included. ...................... 106

Figure17 6.2.5 Map showing geographical boundaries used to define the mid-Atlantic and south-Atlantic models used in the assessment. ............................................................... 107

Figure 186.2.1.1. Comparison of Standardized estimates [( obs-mean)/std.dev] for the three major indices used in the Mid-Atlantic model. .............................................................. 108

Figure 196.2.1.2. Comparison of Standardized estimates [( obs-mean)/std.dev] of the three major indices used in the mid-Atlantic model to the VIMS spring juvenile index and North Carolina Indices. .................................................................................................... 109

8

Figure20 6.2.1.3. Comparison of Standardized estimates [( obs-mean)/std.dev] for the two major indices used in the South-Atlantic model with other available indices for the region. ................................................................................................................................. 110

Figure 216.2.2.1 Proportion of commercial landings by size . .................................................. 111

Figure22 6.2.2.2. Proportion of commercial landings by estimated age (1973?-2002) ........... 111

Figure 236.2.2.3. Predominant size range of Atlantic croaker in the commercial landings overlaid on combined age-length data. ........................................................................... 112

Figure 246.2.2.4. Proportion of recreational landings by size (1981-2002) ............................. 112

Figure25 6.2.2.5 Proportion of recreational landings by estimated age (1981-2002) ............. 113

Figure 266.2.2.6. Predominant size range of Atlantic croaker in the recreational landings overlaid on combined age-length data. ........................................................................... 113

Figure 276.2.2.7. Proportion of SEAMAP catches by size class ............................................... 114

Figure28 6.2.2.8. Proportion of SEAMAP catches by age ........................................................ 114

Figure 296.2.2.9. Predominant size range of Atlantic croaker in the SEAMAP catch overlaid on combined age-length data. .......................................................................................... 115

Figure 306.2.2.10. Proportion of NMFS survey catches by size class ...................................... 115

Figure 316.2.2.11. Proportion of NMFS survey catches by estimated age .............................. 116

Figure 326.2.2.12. Predominant size ranges of Atlantic croaker in the NMFS trawl catch overlaid on combined age-length data. ........................................................................... 116

Figure33 6.2.2.13. Maximum likelihood profile for the prior distribution for the steepness parameter, h, from the covariate analysis of Myers et al.(2002) ................................. 117

Figure 347.1.1. Observed and predicted commercial landings from base Mid-Atlantic model ............................................................................................................................................. 118

Figure 357.1.2. Observed and predicted recreational landings from base Mid-Atlantic model ............................................................................................................................................. 118

Figure 367.1.3. Observed and predicted commercial landings from base South-Atlantic model ............................................................................................................................................. 119

Figure37 7.1.4. Observed and predicted recreational landings from base South-Atlantic model .................................................................................................................................. 119

Figure 387.1.5. Observed and predicted estimates for the NMFS trawl survey for the base mid-Atlantic model ........................................................................................................... 120

Figure 397.1.6. Observed and predicted estimates for the MRFSS index for the base mid-Atlantic model ................................................................................................................... 120

Figure40 7.1.7. Observed and predicted estimates for the SEAMAP index for the base mid-Atlantic model ................................................................................................................... 121

Figure 417.1.8. Observed and predicted estimates for the MRFSS index for the base south-Atlantic model ................................................................................................................... 121

9

Figure 427.1.9. Observed and predicted estimates for the SEAMAP index for the base south-Atlantic model ................................................................................................................... 122

Figure 437.2.1.1. Fully recruited fishing mortality estimates for Atlantic croaker from the base mid-Atlantic model. .................................................................................................. 122

Figure447.2.1.2. Fully recruited fishing mortality estimates for Atlantic croaker from the base South-Atlantic model. ............................................................................................. 123

Figure 457.2.2.1. Spawning Stock Biomass and Age 0 estimates for Atlantic croaker from the base mid-Atlantic model. .................................................................................................. 123

Figure 467.2.2.2. Spawning Stock Biomass and Age 0 estimates for Atlantic croaker from the base south-Atlantic model. ............................................................................................... 124

Figure 47487.4.1. Probability profiles used for the deterministic estimates evaluated in the Mote-Carlo sensitivity analysis. For steepness, see Figure 6.2.2.13. ........................... 125

Figure 497.4.2. Distribution of commercial (A) and recreational (B) fishing mortality rates per year determined using 1,299 Mote-Carlo trails. ...................................................... 126

Figure50 7.4.3. Distribution of spawning stock biomass estimates determined using 1,299 Mote-Carlo trails. .............................................................................................................. 127

Figure51 7.4.4. Distribution of Age 0 estimates determined using 1,299 Mote-Carlo trails. . 128

Figure52 8.1.1. Phase plot of the ratio of F2001/ Fmsy with SSB2001/SSBmsy for the Monte-Carlo simulations. ........................................................................................................................ 128

Figure 538.1.2. Phase plot of the ratio of F avg1999-2001/ Fmsy with SSBavg 1999-2001/SSBmsy for the Monte-Carlo simulations. ................................................................................................. 129

Figure 548.1.3. Estimated fishing mortality rates from the base mid-Atlantic model relative to proposed benchmarks. ................................................................................................. 130

Figure 558.1.4. Estimated spawning stock biomass from the base mid-Atlantic model relative to proposed benchmarks. ................................................................................................. 130

Figure 568.2.1 57Beverton and Holt stock recruitment curve and stock recruit scatter for the base mid-Atlantic model. Vertical line represent SSB(MSY) ....................................... 131

Figure58 8.2.2 Beverton and Holt stock recruitment curve and stock-recruit scatter for the base south-Atlantic model. Vertical line represents SSB(MSY) ................................... 132

Figure 598.3.1 Yield per recruit and spawning potential ratio curve for the base mid-Atlantic model (m=0.3). Avg represents average SPR from 1999-2002 ...................................... 133

Figure 608.3.2 Yield per recruit and spawning potential ratio curve for the base south-Atlantic model (m=0.3). Avg represents average SPR from 1999-2002 ....................... 133

10

Terms of Reference

1. Evaluate adequacy and appropriateness of fishery-dependent and fishery- independent data used in the assessment (i.e. was the best available data used in the assessment). 2. Evaluate adequacy, appropriateness, and application of models used to assess the species and to estimate population benchmarks. 3. Evaluate adequacy and appropriateness of the Technical Committee's recommendations of current stock status based on biological-reference points. 4. Develop recommendations for future research for improving data collection and the assessment. 5. Prepare a report summarizing the peer review panel's evaluation of the stock assessment. (Drafted during the Review Workshop; Final report due October 23, 2003.) 6. Prepare a summary stock status report including research recommendations. (Drafted during the Review Workshop, Final report due October 23, 2003.)

1.0 Introduction

The Fishery Management Plan (FMP) for Atlantic croaker, adopted in 1987, included the states from Maryland through Florida. After a review of early results of the Interstate Fisheries Management Process, the Atlantic States Marine Fisheries Commission (ASMFC) determined that the plan for Atlantic croaker should possibly be revised. . A Wallop-Breaux grant from U.S. Fish and Wildlife Service provided fiscal support for a workshop for this species as well as spot. The results would provide the foundation for a major amendment to the 1987 FMP. The October 1993 workshop at the Virginia Institute of Marine Science was attended by university and state agency representatives from six states. Presentations on fishery-dependent and fishery-independent data, population dynamics and bycatch reduction devices were made and discussed. The results and a set of recommendations were included in the workshop report (ASMFC 1993). Subsequent to the workshop and independent of it, the South Atlantic State/Federal Fisheries Management Board of the ASMFC reviewed the status of several plans to define those compliance issues to be enforced under the Atlantic Coast Fisheries Cooperative Management Act (ACFCMA). The Board found the Atlantic Croaker FMP was vague and no longer valid; they recommended an amendment to define management measures necessary to achieve the goals of the FMP. In the final schedule for compliance under the ACFCMA, the Interstate Fisheries Management Program (ISFMP) Policy Board adopted the finding that the current Atlantic Croaker FMP does not contain any management measures that states are required to implement (ASMFC 2002).

11

A Technical Committee appointed in 1997, compiled data during the summer of 1998. This was the first step in the preparation of a stock assessment. The proceedings of the 1993 workshop as well as data collected by the states and federal agencies since then provided the basis for an amendment to the plan (ASMFC 2002). However, no amendment has been drafted, to date.

2.0 Life History

General Information Atlantic croaker (Micropogonias undulates Linnaeus) occur in coastal waters from the Gulf of Maine to Argentina (Lee et al. 2001). Although not common north of New Jersey, this species is one of the most abundant inshore demersal fish of the Atlantic Coast of the United States (ASMFC 1987). The Atlantic croaker is an opportunistic bottom-feeder on benthic epifauna and infauna and consumes a variety of invertebrates, including polychaetes, mollusks, ostracods, copepods, amphipods, mysids, and decapods, and occasionally fish (ASMFC 1987). Differences in spatial and temporal distribution, as well as differences in feeding behavior, reduce competition between juvenile sciaenids, such as Atlantic croaker and spot, and allow them to coexist in the same area (both spot and Atlantic croaker frequently co-occur in the same habitats – including juveniles). Predators of Atlantic croaker are larger piscivourous species such as striped bass, southern flounder, bluefish, weakfish, and spotted seatrout (ASMFC, 1987). Larvae have been collected from near the edge of the continental shelf to within estuaries of the Mid- and South Atlantic coast (ASMFC 1987). Croaker larvae move from offshore spawning grounds to estuarine areas by mechanisms that are not well understood, but are likely influenced by both behavior of the larvae and physical processes (Barbieri et al. 1994a). Recruitment of young-of-the-year (YOY) croaker to estuarine areas occurs over an extended period of time. Movement into the nursery areas generally peaks in the fall north of Cape Hatteras, North Carolina and in the winter and early spring to the south. Young –of-the-year were collected in October in the Delaware River, October to February in a Virginia Atlantic coast estuary, and July to November in Chesapeake Bay. Recruitment of early life stages to estuaries south of Chesapeake Bay took place from August to April with maximum ingress in December through February for North Carolina, South Carolina, Georgia and Florida (ASMFC 1987). Early life history stages of Atlantic croaker exhibit ontogenetic shifts in prey items and habitat preferences. Larval and post larval Atlantic croaker are primarily zooplantivourous, while detritus appears to be a major component of the juvenile diet. The detritus may be a result of the foraging on benthic infauna and epifauna rather than a source of energy. Post-larval and very young Atlantic croaker occupy estuarine nursery areas, where they are often associated with the shallow marsh habitat over a broad range of estuarine salinities (ASMFC 1987). Temperature induced winter mortality may be an important factor limiting recruitment in the mid-Atlantic bight. Lankford and Targett (2001) determined winter water temperatures at or below 3o C drastically reduced survival of YOY Atlantic croaker. Laboratory experiments

12

indicated 0% survival at 1o C and 1.3% survival at 3o C. There was a size-dependent factor where smaller individuals survived at higher rates than larger individuals (Lankford and Targett, 2001). 2.1 Age Initial studies of the age of Atlantic croaker in the Gulf of Mexico were based on the analysis of marks on scales (White and Chittenden 1977). These researchers found few age groups and concluded that this species has a short life span, early age at maturity and could withstand considerable exploitation. Barger (1985) found that transverse sections of sagittal otoliths gave the most repeatable age estimates of Atlantic croaker from the Gulf of Mexico. Marginal increment analysis indicated that a single mark was deposited annually on the sagittae. Also, eight age groups were found suggesting that scales underestimate the true age of the fish in that area. Ross (1988) aged Atlantic croakers from North Carolina waters also by scale analysis. Subsequently, Barbieri et al. (1994b) used sections of sagittae to age fish from the Chesapeake Bay during 1988-1991. A single annulus formed each year during April and May for all age classes (8); precision of the estimates was very good (99%). Their maximum age was 8 years from Chesapeake Bay collections (Barbieri et al. 1994b). Since this study, the population has expanded and maximum observed age has increased to 12 from fishes landed in Virginia and North Carolina in 2001 (Bobko et al. 2003 and NCDMF 2002). Sections of Atlantic croaker otoliths removed from archeological excavations near St. Augustine, Florida indicated that coastal Indians from the First Spanish period captured fish with a maximum age of 15 years (Hales and Reitz 1992). Since Atlantic croaker have an extended spawning season and recruit to the estuarine nursery areas over an extended period, there are some problems associated with the assignment of ages to fish taken along the Atlantic coast of the U.S. As previously stated, the fish may move into the estuaries north of North Carolina as early as July. This would result in these croakers being approximately seven to ten months of age during their first spring. Along the southeast coast (North Carolina and south), most Atlantic croaker recruit to the estuaries from January through March. These fish would be from two to five months of age during their initial spring. The YOY north of Cape Hatteras form a rather indistinct mark near the core of the otolith that has been designated as the first annulus by some researchers, e.g., Barbieri et al. (1994b). The problem lies in the fact that this mark is not seen in the transverse sections of the sagittae of all fish. In those fish with the ring proximate to the core, the indistinct mark is designated as the first annulus. If the mark is absent and the distance to the first well-defined increment is relatively large, one is added to the number of annuli. South of Chesapeake Bay, some fish do have the hazy area near the core, but many fish lack it. Ages of the fish from North Carolina and south have been determined by designating the first well defined, distinct ring as the first annulus. The ages may be made comparable by either subtracting one from the northern estimates by Virginia researchers or adding one to the counts from North Carolina and South Carolina biologists.

13

It should be noted that a workshop is planned by South Carolina Department of Natural Resources to attempt to standardize procedures for aging east coast sciaenids, including the Atlantic croaker. Age data were available from five sources, and all surveys used sectioned otoliths to age specimen. These were as follows: (1) the Virginia Marine Resources Commission/ Old Dominion University (Virginia commercial landings from 1998-2002); (2) The North Carolina Division of Marine Fisheries (NCDMF) (fishery independent and dependent sources from 1996-2002), (3) Virginia Institute of Marine Science Age and Growth study (1998-2000); (4) South Carolina DNR aging of SEAMAP samples (2001-2002); and from commercial landings from Maryland (1999-2001) (Tables 2.1.1 and 2.1.2).

2.2 Growth

The size-at-age for Atlantic croaker is highly variable (Chittenden et al. 1994). Croaker grow rapidly during the first year; but the rate decreases during the second year and remains comparatively low thereafter. Barbieri et al. 1993 found on average, 64% of the cumulative total observed growth in length occurred in the first year and 84% was completed after two years. There was no difference found in the total length-total weight relationship between sexes (Chittenden et al. 1994).

Given the uncertainty associated with ages determined using scales, only those from sectioned otoliths were included. Growth parameters derived from the available otolith age data were compared with the literature (Table 2.2.1). The increased number of older fish in recent samples from Virginia and North Carolina result in larger von Bertalanffy estimates of theoretical maximum size (L∞) than those of Barbieri et al. 1994a. Estimates of the growth parameter, K appear to be lower for data in the recent time series than those of Barbieri (1994a). There also appears to be a similarity in the recent von Bertalanffy estimates to those estimated by Hales and Reitz (1992) for Atlantic croaker from archeological sites (Table 2.2.1).

2.3 Reproduction Atlantic croaker are multiple spawners with asynchronous oocyte development and indeterminate fecundity. At a population level, spawning extends over a six month period (July- December, could extend into January). Some authors suggest that individual fish spawn for only 2-3 months (Chittenden et al. 1994). Atlantic croaker spawn in the lower Chesapeake Bay as well as in coastal oceanic waters (Chittenden et al. 1994). Apparently, spawning starts in Chesapeake Bay and continues offshore and south as Atlantic croaker migrate out of the estuary. However, the occurrence during the fall of some regressing and resting females in Chesapeake Bay indicates that at least some individuals may complete their spawning in estuarine waters. A re-examination of the historical ichthyoplankton studies of the Chesapeake Bay would provide an indication of the magnitude of estuarine spawning for this species.

14

2.3.1 Sex ratio Atlantic croaker in the Chesapeake Bay region showed temporal changes in sex ratio (Chittenden et al. 1994). In 1990-1991, Chittenden et al. 1994 found the contribution of the males in the Chesapeake Bay decreased at the beginning of the spawning season (June-July) and reached a minimum in September-October Males became more abundant again during November-December. Between 1989 and 2002 the annual proportion of females for the Virginia commercial fishery range between 0.54 and 0.8 with an average of 0.67.

2.3.2 Size and Age at Maturity Based on samples of the commercial catches in the Chesapeake Bay and the Virginia and North Carolina coastal waters (n = 3091) during 1990 to 1991, Barbieri et al. (1994b) determined that Atlantic croaker mature at a small size and early age. Males and females started to mature at 170 and 150 mm total length, respectively. At larger sizes, the percentages of mature fish in the samples increased rapidly. Estimated mean length at first maturity was 182 mm TL for males and 173 mm TL for females. All individuals greater than or equal to 250-260mmTL were mature, regardless of sex. They also indicated that the same general pattern held for the maturity schedule by age. More than 85% of both males and females were sexually mature by the end of their first year. 2.4 Stock definitions Genetic population structure in Atlantic croaker (Micropogonias undulatus) was examined by using the polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analysis of mitochondrial DNA (mtDNA) (Lankford et al. 1999). Juvenile croaker from three U.S. Atlantic localities (Delaware, North Carolina, and Florida) and one Gulf of Mexico locality (Louisiana) were screened to document the magnitude and spatial distribution of mtDNA variation in this species. The objectives were to evaluate the integrity of Cape Hatteras, North Carolina, as a genetic stock boundary; and to estimate levels of gene flow among Atlantic localities to provide an improved basis for future decisions regarding coastwide management of this fishery resource (Lankford et al. 1999). There was significant heterogeneity between Atlantic and Gulf of Mexico samples, suggesting restricted gene flow between these two regions. Analysis of molecular variance also indicated regional (Atlantic versus Gulf) population structure, but provided no evidence that Cape Hatteras represents a genetic stock boundary. These findings are consistent with: 1) a single genetic stock of M. undulatus on the Atlantic coast, and 2) separate, weakly differentiated stocks in the Atlantic and Gulf of Mexico (Lankford et al. 1999)

15

3.0 Fishery Description

3.1 Brief overview of Fisheries Earlier records of commercial landings exist for some states, but because the data are incomplete we used records from 1950 to the present in this report. North Carolina commercial landings were low throughout the 1950s in comparison to Virginia. Sustained levels of high landings (greater than 5 million pounds) occurred in North Carolina from 1974-1990 and 1995-2001 and in Virginia from 1954-1959, 1976-1978, and 1993-2002. The recreational catch statistics collected by the National marine Fisheries Service (Marine Recreational Fishery Statistics Survey, MRFSS) provided the data for the recreational landings component of this assessment. Data from 1981 through 2002 was used in. Recreational landings of Atlantic croaker (Type A + B1 in numbers), from Massachusetts through the Atlantic coast of Florida, have varied between 2.8 million fish (1981) and 13.2 million fish (2001), with landings showing a strong linear increase over this period (Table 5.2.2.1). 3.1.1 Commercial Fishery Commercial landings of Atlantic croaker varied from one million pounds in 1970 to nearly 30 million pounds in 1976 and 1977 along the Atlantic coast. From 1996 to 2001, commercial landings have exceeded 20 million pounds annually. Annual landings consistently increased from a low of 3.7 million pounds in 1991 to 27 million pounds in 1997 (Table 3.1.1.1). North Carolina landings have continued to grow since 1993, to a maximum in 2001. However, the largest increase in landings was in Virginia, where only 164,000 pounds were reported in 1991, and more than 12 million pounds have been landed annually in Virginia since 1997. Coastwide landings of Atlantic croaker have remained steady at 25 to 28 million pounds from 1997 to 2001. This species is a major component of the commercial catches of Virginia, North Carolina, and Maryland. Gill nets, haul seines, trawls and pound nets accounted for most of those landings. (ASMFC 2002). In 2001, the total commercial value of croaker landings was $7,274,111(Table 3.1.1.2). Atlantic croaker is the major component of the North Carolina “scrap fishery”. A number of regulations instituted by North Carolina, (the elimination of flynet fishing south of Cape Hatteras (1994); the introduction of BRDs in shrimp trawls (1992, by proclamation authority); limits on the incidental catch of finfish by shrimp and crab trawls in inside waters (since 1970s); and culling panels in long haul seines (1999) may have indirectly reduced catches of juvenile croaker and changed the size and age distributions of the harvest. In Georgia, trawl-caught croaker is sold as unsorted mixed fish along with spot, whiting, and small flounder; therefore, commercial landings are a tenuous measurement there. Small Atlantic croaker were previously a major part of the bycatch of the south Atlantic shrimp trawl fishery, however the use of TED’s and BRD’s has reduced this bycatch by an unquantifiable amount (ASMFC 2002).

16

3.1.2 Recreational Fishery Between 1981 and 1990 annual average recreational landings (in numbers) amounted to 6.0 million fish, while more recently, between 1997 and 2002, recreational landings have ranged from a minimum of 9.1 million fish to a maximum of 13.2 million fish with average annual landings of 10.8 million fish. The increased landings in recent years have been at the northern range of the fishery (Massachusetts to North Carolina) particularly in New Jersey, Delaware, Maryland, and Virginia (Figures 5.2.2.1 and 5.2.2.2, Tables 5.2.2.1 and 5.2.2.2). During the past 10 years, recreational landings in Virginia, accounted for an average of 69 and-67 % of the total landings in numbers and weight, respectively. Landings from states north of Delaware accounted for sporadic and negligible landings of Atlantic croaker. Recreational landings at the southern range of the fishery (South Carolina through the Atlantic coast of Florida) have remained relatively stable, since 1997, with an annual average of 4.6 million fish and 2.5 million pounds (1997 – 2002). Recreational landings from the southern range of the fishery accounted for approximately 4.5 % of annual coast-wide landings between 1997-2002. The majority of landings in the southern region of the fishery were made on the Atlantic coast of Florida (Tables 5.2.2.1 and 5.2.2.2)

3.2 Regulations and Management History

The 1987 FMP for Atlantic croaker identified the following management measures for implementation: 1) Promote the development and use of bycatch reduction devices through demonstration and application in trawl fisheries. 2) Promote increases in yield per recruit through delaying entry to croaker fisheries to age one and older. Although the ISFMP Policy Board judged that the FMP management recommendations were too vague and did not furnish objective compliance criteria, progress has been made on developing bycatch reduction devices (BRD’s). The October 1993 workshop proceedings summarized experimental bycatch reduction work and examined the implications of bycatch reduction on the populations of Atlantic croaker and spot (ASMFC 1993). It was clear that there were economically viable shrimp gears that reduce finfish bycatch. North Carolina closed ocean waters south of Cape Hatteras to the South Carolina state line for flynets in 1994. These actions may indirectly affect the fishing impact on croaker (ASMFC 2002). Table 3.2.1 summarizes the current state regulations for Atlantic croaker. Currently no regulations directly govern fishing practices for Atlantic croaker in North Carolina. However, the regulation, limiting the scrapfish catch to 5,000 pound per vessel per day, has an indirect effect since Atlantic croaker comprise a large percentage by weight landed by NC commercial fishing gears. BRDs were required in all North Carolina shrimp trawls in the fall of 1992 by proclamation. Restrictions such as a minimum mesh size (3” square or 3.5” diamond) in 1991 and the closure of ocean waters south of Cape Hatteras to flynets in 1994, also moderated the

17

exploitation of croaker. Initial studies in long haul seines in 1996 produced a reduction in the average catch of the scrap fish species. The NCDMF adopted a permanent rule in April 1999 to require escape panels in long haul seines in the southern areas of the state. Some preliminary work has been done with sciaenid pound net fishermen along the Outer Banks to test a similar panel design in this fishery. A reduction of sub-adult croaker harvested should increase both spawning stock biomass and yield per recruit. The Potomac River Fisheries Commission promotes the use of large mesh bycatch reduction panels in all pound nets, but use is voluntary (fishermen who use the escape panels are allowed to keep a by-catch of weakfish). It is estimated that the panels allow the release of 100% of captured croaker below the minimum legal size of nine (9) inches (ASMFC 2002). The states of Florida through North Carolina have promoted and required the use of TED’s (turtle excluder devices) and BRD’s for trawls in state waters. Direct finfish trawling in inside estuarine waters has been banned in North Carolina since 1931. Finfish bycatch limits have been set since 1970s for non-finfish targeting trawls (i.e.: shrimp and crab) in inside estuarine waters and presently allows for only 500 pounds of finfish from December 1 to February 28 and 1,000 pounds of finfish from March 1 to November 30. North Carolina has implemented minimum stretch mesh size restrictions in shrimp trawls (1 1/2" tailbag) and crab trawls (to take hard crabs-3"; to take soft or "peeler" crabs-2") since 1991. Ocean trawls or flynets in ocean state waters have a minimum stretch mesh size since 1997 (4” main body, 3” extension, and 1 3/4” tail bag). Florida has a maximum shrimp trawl size. A ban on trawling in Virginia waters has been in effect, since 1989. Before you target a reduction, you need to measure the magnitude of the bycatch of this species, presently we lack good estimates for many of the South Atlantic states. Size limits that are in place in the states have been there for several years and do not represent a response to the FMP. In order to minimize recreational discard mortality, a new amendment may evaluate the concept of encouraging the use of hook types, which minimize such mortality (ASMFC 2002).

4.0 Habitat Description

The estuarine nursery areas for Atlantic croaker populations differ considerably among locations, possibly in response to tidal range. Where the range is less than 0.5m (20 inches), shallow open water areas at the landward extremities of large bays, as are shallow creeks, ponds, and lakes intimately associated with marsh are of major importance to juveniles. Where the tidal influence is stronger, large numbers of small juveniles have been collected from small tidal streams in the spring (ASMFC 1987); however, most reports indicate shallow areas are avoided and juvenile croakers are concentrated in the deep, main channels of estuaries as in the Delaware River, Chesapeake Bay, and the Cape Fear River. Apparently, shallow areas become less suitable for juvenile croakers as daily fluctuations of water level increase (ASMFC 1987). Atlantic croakers are eurythermal with the early life stages more cold tolerant than adults. Juvenile croakers have been caught at water temperatures ranging from 0o to 32oC (ASMFC 1987). Atlantic croaker are also euryhaline being taken at salinities from 0 ppt to 70 ppt. More juveniles are associated with salinities in the oligohaline and mesohaline range (0.5 to 18 ppt)

18

and the 0 to 70 ppt are the extremes (ASMFC 1987). As Atlantic croaker grow, they are much more likely to be found at high salinities (ASMFC 1987).

5.0 Data Sources

5.1 Commercial

Commercial landings data were taken from NOAA general canvas reports for all states, including the east coast of Florida. No observer data were available to quantify discard levels.

Biological samples were from state surveys. Age, length, and weight data of the Atlantic croaker commercial fishery have been sampled at fish houses by NCDMF since 1982, and VMRC since 1989. Maryland DNR has had a pound net survey in Chesapeake Bay since 1993. Limited age and weight data from MD are available since 1999.

5.1.1 Data Collection Methods

5.1.1.1 Survey Methods

NCDMF The NCDMF has sampled major commercial fisheries since 1982. Atlantic croaker were sampled by gear, market category (in culled catches only), and area fished at local fish houses. Fishes were measured to the nearest mm (TL) and sample weights as well as total weights were taken to expand the sample data to the entire catch. Beginning in 1994, NCDMF instituted a trip ticket system to track commercial landings. Total catch by gear, area and market category were used to expand these data. To obtain overall annual distributions or mean landed CPUE in a fishery, the expanded values were weighted by the tri-annual commercial landings of the respective fishery in order to account for seasonal and between fishery differences in the magnitude of the landings. In 1994, the landings collection method changed from a voluntary dealer reporting system to a mandatory trip ticket system. Therefore, data may not be comparable between pre-1994 and post-1994 landings. Scrapfish sampling was initiated in 1986. Total weight of a species in the scrapfish samples was calculated by determining the proportion of a species in the subsample and expanding that to the respective species proportional weight of the total scrapfish for the trip. The number of individuals per species in the scrapfish component was calculated by expanding the number of individuals in the sample to represent the total weight of the species for the scrapfish in the samples. Estimates of scrapfish landings for individual species were determined by applying the tri-annual ratio of marketable fish to scrapfish in the fish house samples to the reported tri-annual marketable landings. These trends are only from 1986 on, due to the lack of bait sampling prior to 1986. Sub samples of Atlantic croaker were purchased to excise otoliths for age determination across the major commercial fisheries.

19

VMRC At seafood dealers and buyers, commercially caught Atlantic croaker were sampled from 50-pound boxes of the graded catch. These were measured ( mm TL) and weighed (0.1 lb). Market category, harvest area, gear type and total catch were noted. Beginning in 1999, samples were purchased to excise otoliths for age determination. All aging studies (processing and reading) were done at Old Dominion University’s Center for Quantitative Fisheries Ecology.

MD DNR Since 1993, commercial pound nets were surveyed during June through September. Atlantic croaker were sampled for length data. Beginning in 1999, limited age, sex and weight data were collected. All otoliths were processed and read by SC DNR.

5.1.1.2 Sampling intensity Sampling intensity, relative to the magnitude of the catch, was generally low especially in MD, VA, and NC in the late 1990’s as harvests increased. No other length data were available from other states. Sampling for length from the MD commercial fishery ranged from 0.7 lengths/mt in 2000 to 10.8 lengths/mt in 1994 (average 1993 – 2001 = 3.9 length / mt). Virginia’s sampling intensity ranged from 0.47 lengths / mt in 2000 to 63 lengths / mt in 1991 (average 1989 – 2001 = 14.0 lengths / mt). North Carolina’s sampling intensity ranged from 0.2 lengths/ mt in 1977 to 18.5 lengths / mt in 1992 (average 1977 – 2001 = 7.4 lengths / mt).

5.1.1.3 Biases Substantial biases may exist in the Atlantic croaker sampling programs among the states collecting biological data. There are distinct seasonal and gear differences (selectivity) among the fisheries. The rapid growth of Atlantic croaker also makes it necessary to sample the catch throughout the year. Initially, the Stock Assessment Subcommittee attempted to segregate the biological data by trimesters, but available information was inadequate to characterize the commercial catch at such a resolution. The Subcommittee then decided that aggregating the biological data to broader time periods would introduce too much bias.

5.1.1.4 Aging methods North Carolina Atlantic croaker sagittal otolith samples were collected monthly from the winter trawl, long haul seine, pound net, sink net, recreational hook and line fisheries, and NCDMF independent programs. Sagittal otoliths have been collected since 1996. Each month, samples (n=15) are distributed across the size range in 15-mm size classes starting at 100 mm total length. Sagittal otoliths were removed, cleaned and stored dry. Total length to the nearest millimeter, weight to the nearest 0.01 kg, date, gear, and water location were recorded for each sample.

20

A transverse section through the focus on a plane perpendicular to the horizontal axis of the left otolith was prepared using a Hillquist thin-sectioning machine as described by Cowan et al. (1995). The system was calibrated with an ocular micrometer before each reading session. Sections were viewed under reflected light at 21X magnification. Annuli, marginal increment, and otolith size were measured (mm) on an image projected on a high resolution monitor from a video camera mounted on a microscope. Ages were assigned on the number of otolith annuli viewed. Sections were read and annuli measured by the aging lab biologist then independently read by the species lead biologist. Any differences were resolved or the data were not included. A three-year report is compiled for species specific (seasonal based within a calendar year; winter January-March and October-December; summer April-September) age-length keys and applied to expanded length-frequency data to determine length at age for landed catches on an annual basis (NCDMF 2001). South Carolina In the laboratory, the left sagittae were viewed under low magnification with a binocular microscope (10X) and marked with a soft lead pencil on the core. These were then embedded in epoxide resin in silicon molds. After the resin had polymerized, the embedded otoliths were glued to a card held in a jig attached to the arm of a low speed saw. The otolith was positioned so that a transverse section ~0.5-mm thick could be taken through the core. The Isomet Saw was equipped with a pair of diamond wafering blades, separated by a plastic washer so that the section could be taken with a single cut. The resulting section was mounted on a labeled microscope slide with Cytoseal-XLY. After polymerization of the mounting medium, slides were stored in boxes until viewing. These were examined with a Nikon SMZU microscope equipped with a Supercircuits model PC – 23C high resolution camera with transmitted light. The video image was captured by a frame grabber board in a personal computer and was subsequently analyzed with the OPTIMAS® image analysis software. The following measurements were taken on each otolith section: (1) radius – distance in mm from the center of the core to the edge of the section as measured along the sulcus acousticus; (2) a1 – distance in mm from the center of the core to the distal edge of the first annulus; (3) a2 – distance in mm from the center of the core to the distal edge of the second annulus; (4) a3 to an – distance from the center of the core to the distal edge of the third annulus and from the core to the distal edge of the nth annulus; (5) marginal increment – distance from the distal edge of the last annulus to the edge of the otolith section. Some Atlantic croaker otoliths varied with respect to diffuse, undefined marking near the core of the otolith. These diffuse areas were not interpreted as being a ring. We called the first annuli the first well defined opaque band that could be traced around the entire section.

5.1.2 Commercial landings

Atlantic state’s commercial landings of Atlantic croaker exhibited three periods of peak landings: 1955 – 1959, 1975 – 1980, and 1995 - present (Figure 5.1.2.1). The highest landings were in

21

1977 at 13,532 mt. The current period of elevated landings is more than seven years. Conspicuously low levels of harvest were evident during the 1960’s and early 1970’s. Three gear types have historically accounted for 95% of the harvest (Table 5.1.2.1). Haul seine and trawl fisheries accounted for an average of 31% and 35% of total Atlantic croaker harvest since 1950, respectively. Pound net and haul seine fisheries each accounted for an average of 16% of total landings, 1950 – 2001. The commercial harvest has been dominated by NC and VA since 1950 (Table 5.1.2.2). North Carolina averaged 59% of the annual commercial landings among Atlantic coast states since 1950 and Virginia had a mean of 33.2%. Recently, (1997 – 2001) the mid-Atlantic states (NJ, MD, VA) had a higher proportion of the total commercial catch than their historic average. For example, New Jersey landings comprised 6.4% of the coastal catch during 1997 – 2001 (long term average = 1.6%), Maryland averaged 6.0% (long term average = 3.7%) and Virginia averaged 47.3% (long term average = 33.2%).

5.1.3 Commercial discards and bycatch

Quantifying bycatch and discard of Atlantic croaker is difficult. North Carolina maintains a regulated scrap fishery. However, the scrap fishery is not accounted for in the NOAA general canvas data. The incidental catch of Atlantic croakers in the shrimp trawl fishery (predominately NC and south) is a major source of bycatch and discard mortality. Bycatch of Atlantic croaker was estimated at 5.8 mt to 12.7 mt (NC to FL) from 1973 to 1975, and 611 mt and 2283 mt (SC to FL) in 1992 and 1993, respectively (Diamond et al. 1999). Beginning in 1992, BRDS were mandated for the shrimp trawl fishery in North Carolina. No estimates of bycatch were available post-implementation of BRD mandates.

5.1.4 Commercial catch rates Data were insufficient (spatially and temporally) to calculate CPUE from the commercial fishery.

5.1.5 Commercial catch at age

The subcommittee investigated using length and age data to compile a catch at age matrix for Atlantic croaker in four month intervals. However, the quality of the length and age data were not sufficient to complete this task.

5.2. Recreational

Two sources of recreational landings data for Atlantic croaker examined; the recreational catch

22

statistics collected by the National Marine Fisheries Service (Marine Recreational Fishery Statistics Survey, MRFSS) and the National Marine Fisheries Service’s headboat survey. On the Atlantic coast, the headboat survey collects data from headboats operating South of the Virginia-North Carolina border to Florida. However, examination of the data set revealed that species-specific information on Atlantic croaker was not available from the headboat survey.

5.2.1 Data Collection Methods

The MRFSS survey monitors fishing activity by recreational anglers by state, wave (two-month periods), mode of fishing and area fished. The recreational catch statistics from 1981-2002 and related materials were obtained from the MRFSS web site at http://www.st.nmfs.gov/st1/recreational/.

5.2.1.1 Survey Methods

A detailed description of the MRFSS survey methods is available at http://www.st.nmfs.gov/st1/recreational/pubs/data_users/index.html. To summarize, the survey consists of two independent and complementary parts: 1) a random telephone survey of households in the coastal counties of the eastern United State which is used to determine the number of recreational fishing trips conducted by the three modes during two-month time periods. 2) Angler interviews which collect information on the number of fish seen by the samplers, the number of fish that were caught and were unavailable to the sampler (Type B) because the fish could have been eaten, used for bait, or released and information on length and weights of fish available for inspection. The data from the two components are combined with U.S. Bureau of Census data to produce estimates of recreational catch, effort by state, wave, mode of fishing and area fished.

5.2.1.2 Sampling Intensity

The allocation of sampling for the telephone survey within a state is proportionally allocated based on the square root of the number of full-time occupied households in each county (MRFSS, 1999). For the intercept survey, sampling is stratified by state, mode, and wave with a minimum of 30 intercepts per stratum. Samples are allocated beyond the minimum in proportion to a 3-year average of fishing pressure (MRFSS 1999). For the intercept data, the number of interviews conducted in which Atlantic croaker were reported and the number of fish measured are presented in Table 5.2.1.2.

5.2.1.3 Biases

MRFSS estimates are designed to be unbiased and are based on a stratified random sampling design of fishers, which are combined with a random telephone survey. However, potential bias in the estimates could arise if the sampler were to have selected interviewees non-randomly. On occasion, there have been instances when the random telephone survey was found to be unrepresentative and an average estimate of trips has been substituted. Most recently, the 2002

23

telephone survey data were discarded for waves 2 and 3 and effort estimates based on a three- year average (1999-2001) for those waves was used.

5.2.1.4 Biological Sampling

As part of the intercept survey, MRFSS samplers also collect information on length and weights of fish measured. For Atlantic croaker, there is no other biological information from the MRFSS.

5.2.1.5 Aging Methods Not Applicable

5.2.1.6 Development of Estimates

Estimates of landings in numbers and weight (Type A +B1), released landings (Type B2) and the total recreational trips are those published in the MRFSS. The trip estimates are for all recreational trips for the strata (i.e. by state, year wave, mode, area). An estimate of the number croaker trips within the strata was calculated by weighting the MRFSS trip estimates by the proportion of potential croaker trip intercepts to total intercept trips for the strata. See Section 5.2.4 for details on how potential croaker trips were identified. Total intercept trips were the number of unique angler trips for a given strata from the MRFSS intercept data set.

Recreational landings by size were determined using the length measurements from the intercept data set and the MRFSS landings estimates in numbers (Type A+B1). Examination of the data indicated that lengths samples were adequate (at best) to work at the state-year-wave level. Size distributions, based on 10 mm increments at the state-year-wave level were applied to the landings and released landings separately. However, there were many cells that had fewer than 50 length measurements per cell. For those cells that had less than 50 measurements, a size distribution based on a collapsed group of cells was used in a hierarchical manner. The levels of “collapsed length distributions ” were:

1) If the number of length measurements were 50 or greater, those lengths were used to represent the state-year-wave cell. 2) If the number of length measurements were < 50, the size distribution applied to the cell were based on state-year- wave group. Two wave groups used were; waves 1 to 3 collapsed and waves 4 to 6 collapsed. 3) If, after using the collapsed size distribution the number of measurements for the cell was < 50, the length distribution used to fill the cell was based a size distribution at the state-year level. 4) If, after using the previous collapsed size distribution, the sample size was < 50, a size distribution based on measurements at a region-year level were applied. The fishery was divided into three regions; 1) Northeast - Virginia and North. 2) North Carolina and 3) Southeast -South Carolina and south. After using this final criteria there were a small number of cells (4) with less than 50 measurements, which were not collapsed further.

24

Once the landings were assigned a size distribution, the type A+B1 landings were appropriately apportioned among the size ranges representing the cell.

For the stock assessment, recreational discard mortality was estimated at 10% of Type B2 (released fish) estimates by numbers. Given the lack of information on discard mortality estimates, the estimate used was based on a consensus of the stock assessment-working group, but may not represent the true discard rate for the fishery. Recreational discards are those fish caught and released alive, but assumed to die as result of such factors such as hooking mortality and improper handling. As there are no weight estimates for recreational discards, the weighted size distribution of released estimates was used with a length-weight relationship (see section 6.2.2) to estimate the weight of the recreational discards. .

In order to determine the weight of the discards, a weighted size distribution for the Type B2 estimates was required. As there was no information on the size of released fish (Type B2) three approaches to assigning a size distribution were evaluated. The first approach was based on assigning a size distribution similar to those assigned for the Type A+B1 landings. The second approach used the median size class for a given cell and using the length-weight relationship determined the landings for the cell. Since there are no size regulations on Atlantic croaker, except in Maryland and Georgia (PRFC has no size limit but does have a 25-croaker limit), the third approach was based on the assumption that released fish are likely to be representative of the lower range of the size distribution of those fish measured. The size at the 10, 15,20, 25 and 50th percentiles of fish measured within a cell was used to truncate the size distribution used for the released landings (Table 5.2.1.6). The truncated size distributions were used to determine the weight of the discards using the length-weight relationship.

5.2.2 Recreational Landings

From 1981-2002, recreational landings of Atlantic croaker (Type A+B1 in numbers), from Massachusetts through the Atlantic coast of Florida, have varied between 2.8 million fish (1981) and 13.2 million fish (2001), with landings showing a strong linear increase over this period (Table 5.2.2.1). Between 1981 and 1990 annual average recreational landings (in numbers) amounted to 6.0 million fish, while more recently, between 1997 and 2002, recreational landings have ranged from a minimum of 9.1 million fish to a maximum of 13.2 million fish with average annual landings of 10.8 million fish. The increased landings in recent years have been at the northern range of the fishery (Massachusetts to North Carolina) particularly in New Jersey, Delaware, Maryland, and Virginia (Figures 5.2.2.1 and 5.2.2.2, Tables 5.2.2.1 and 5.2.2.2). During the past 10 years, recreational landings in Virginia, accounted for an average of 69 and-67 % of the total landings in numbers and weight, respectively. Landings from states north of Delaware accounted for sporadic and negligible landings of Atlantic croaker. Recreational landings at the southern range of the fishery (South Carolina through the Atlantic coast of Florida) have remained relatively stable, since 1997, with an annual average of 4.6 million fish and 2.5 million pounds (1997 – 2002). Recreational landings from the southern range of the fishery accounted for approximately 4.5 % of annual coast-wide landings between 1997-2002. The majority of landings in the southern region of the fishery were made on the Atlantic coast of

25

Florida (Tables 5.2.2.1 and 5.2.2.2). The precision of recreational landings is expressed as proportional standard error (PSE), which describes the standard error of the estimate relative to the estimate (MRFSS 1999). MRFSS (1999) noted that PSE estimates less than 20% are commonly observed for commonly caught sport fishes. For Atlantic croaker, the PSE by state varies over the time series, with the major mid-Atlantic states (North Carolina and Maryland) being associated with PSE values between 8 –12 % (Table 5.2.2.3). Estimates of Atlantic croaker landings from the south-Atlantic states (South Carolina, Georgia and Florida) were associated with PSE values between 15-30% in recent years (Table 5.2.2.3).

Atlantic croaker were primarily caught in inland waters by fishermen in private or rental boats or fishing from the shore. Fishermen in private/rental boats fishing in inland waters represent on average 71% of landings by numbers since 1993 (Table 5.2.2.4). Landings from offshore waters account for a small portion of the recreational landings (Figure 5.2.2.3). Private/rental boats accounted for the majority of Atlantic croaker landings (Figure 5.2.2.4). During the early to mid 1980’s, shore fishing accounted for a relatively large portion of the landings. However, more recently landings by shore fishers and charter/party boats have remained low and stable (Figure 5.2.2.4). Recreational fishing for Atlantic croaker occurs mostly in the summer (Figure 5.2.2.5) and the majority of landings take place in waves 3 and 4 (May-August). However, at the southern range of the fishery, the landings occur over a slightly extended period from May-October.

Between 1981 and 2002 total recreational effort in state-wave-mode-area combinations where Atlantic croaker were caught has increased in a linear trend from a low of 7.7 to a high of 25 million trips (Table 5.2.2.5; Figure 5.2.2.6). Total recreational effort in the northern range of the fishery (North of North Carolina) accounted for on average 59% of total trips between 1997 and 2002. Estimates of targeted Atlantic croaker trips also show a linear increase between 1981 and 2002 from a low of 3.0 to a high of 14.2 million trips (Table 5.2.2.5; Figure 5.2.2.6). The majority of targeted croaker trips occurred in the northern range of the fishery with an annual average of 80% of total trips in the fishery in recent years (1997-2002).

The size distribution of Atlantic croaker weighted by Type A+B1 landings indicate that at the northern region of the fishery, an increase in the modal size in recent years was evident (Table 5.2.2.6) with the median size increasing from 245 in 1981 to 335 mm TL in 2002 (Figure 5.2.2.7). At the southern range of the fishery the modal size of Atlantic croaker landed has remained relatively stable, fluctuating between 265 and 295 mm TL (Table 5.2.2.7; Figure 5.2.2.8).

5.2.3 Recreational Discards Recreational discards are included in the MRFSS estimates as the number of fish released alive (Type B2). In 1981, Atlantic croaker released by fishermen amounted to 1.2 million fish, and by 2002 the number of released fish had increased almost tenfold (Table 5.2.3.1). At the northern range of the fishery, which accounts for the majority of landings, the ratio of fish released alive

26

to those landed has remained relatively stable over last 10 years, with an average of 1.2 fish being released for every fish that was kept (Figure 5.2.3.1). At the southern range of the fishery, the ratio of Atlantic croaker released to those kept in recent years (1998 onwards) has been similar to those observed for the northern range of the fishery. Prior to 1998, the ratio of fish released to those kept were lower for the southern range of the fishery than those observed for the northern range of the fishery (Figure 5.2.3.1). The estimated numbers and weight of recreational discards based on the different methods used are presented in Table 5.2.3.2.

5.2.4 Recreational Catch Rates

In developing a MRFSS catch rate index for Atlantic croaker, an important factor is defining a sampling unit. Based on the discussions at the data workshop an Atlantic croaker trip was identified using three methods. These were: 1) Original: defined a croaker intercept-sampling unit as one where either croaker was caught

or where the angler recorded croaker as a targeted species, but did not catch any. 2) Jacquard: Used a binary similarity index to identify a suite of species with which Atlantic

croaker were associated, and defined a sampling unit as one where any of those species were caught. This was based on a Jacquard type index (Krebs 1989) and was determined for each state. The species that had the six highest coefficients (this included croaker) were used to identify a sampling unit. Jacquard’s index can be defined as:

cba

aS j

Where: a = no of samples where Atlantic croaker and species j was present b = no of samples where Atlantic croaker were present and species j was not present (unique Croaker samples) c= no of samples where species j was present but Atlantic croaker was not present.

3) Strata: identified all state-year-wave-area-mode strata where Atlantic croaker were caught and used all sampling units within those identified strata as potential Atlantic croaker trips. In general, these three methods made changes to the number of zero cells added. Comparing the preliminary results using the three methods indicated that:

1) As Atlantic croakers are not commonly listed as a targeted species, the number of zero samples added to method 1 (original) was relatively small. This is probably the least appropriate method.

2) In states where the species occurred rarely in the early years (e.g. NJ and DE) the strata method cannot add samples. It is dependent on the species being present.

3) For some states, the differences between strata and the jaccard method reflect that within the strata there are different target species groups. However, for some states the strata and jaccard methods provide similar estimates.

27

Statistical Analysis of Catch-Rates

The base data set used for estimating the MRFSS catch rates was determined using the potential croaker sets identified using method 2 (jaccard). Table 5.2.4.1 shows the species included in defining a croaker set for each of the states used in the analysis. Due to small sample sizes, data from Massachusetts, Rhode Island and New York were excluded. The data were further reduced to only include hook and line sets. The response variable used in the analyses was based on total number of Atlantic croaker per trip (Type A+B1+B2). Two statistical models were used to estimate MRFSS catch rates. These were:

1) A general linear model where log (total number of croaker catch +1) was the response variable. Explanatory variables used in the full model were state year, wave area and mode (treated as classes) and hours fished and contributors, which were, treated as continuous explanatory variables. A state by year interaction term was also included in the full model.

2) A generalized linear model using a negative binomial distribution, using a log link was also carried out. The response variable was the number of croaker per trip (A+B1+B2). The explanatory variables used were similar to the general linear model. However, as the model would not converge within the allotted trials, a state by year interaction term-was not included.

Preliminary evaluation of both statistical models revealed that all explanatory variables were statistically significant (P < 0.01). Given the significant year by state interaction term for the log transformed GLM, both models was re-run by state. A comparison of the normalized catch rates by state revealed that the fishery could be broadly categorized into a northern and southern region. These two groups were the region Virginia and North and the region South Carolina and south. Catch rate trends in North Carolina were intermediate to those seen for the northern and southern regions. For the northern states, the recent time trend indicates higher catch rates than normal while for the southern states, catch rates appear to be fluctuating around or just below their normal levels in recent years. Based on an evaluation of abundance trends in the fishery independent indices, participants at the stock assessment workshop (see section 6.2), concluded that developing separate MRFSS indices for the northern and southern range of the fishery was the most appropriate approach to evaluating catch rates. As such, the data were partitioned into the mid-Atlantic (North Carolina, Virginia, Maryland, Delaware and New Jersey) and the south-Atlantic (South Carolina, Georgia and the Atlantic coast of Florida) and analyses were conducted separately using the protocol described for the preliminary evaluation. Back transformed least square means by year were used as estimates of the catch rate.

Summary statistics for the models are presented in Table 5.2.4.2. For the negative binomial generalized linear model, all explanatory variables were statistically significant (P < 0.01). As such, a reduced model was not developed. For the general linear model, explanatory variables included for both regions were statistically significant with the exception of the number of hours fished in the southern model (Table 5.2.4.2).

Catch rates developed using the two statistical models are presented in Table 5.2.4.3. In general

28

the catch trends from the negative binomial generalized linear model and log-transformed general linear model were similar. For the mid-Atlantic region in 1981/82, the log transformed GLM produced negative estimates that were not significantly different from 0. This was in part because of the low number of trips in which Atlantic croaker were caught. For the northern range of the fishery, catch rates appear to have steadily increased over the time series with a peak in 1999 (Figure 5.2.4.1). For the southern region catch rates appear to be more stable over the time series, except between 1990-1992, when the catch rates were much higher (Figure 5.2.4.2).

5.2.5 Recreational Catch-at-Age

No information.

5.3 Fishery- Independent Survey data For this analysis eight fishery independent surveys were available. An inspection of the fishery independent indices revealed that they primarily targeted juveniles to Age 1, though older age classes were evident in some indices (NMFS and SEAMAP). The eight fishery independent indices were the NMFS fall trawl survey, SEAMAP trawl survey, VIMS trawl survey, North Carolina DMF juvenile estuarine and sound surveys, Maryland DNR juvenile index, and the Florida FWC fishery independent trawl and seine surveys. However, of the available indices the NMFS trawl survey and SEAMAP indices were identified for use in this assessment (see section 6.2.1) together with the possible use of the VIMS trawl survey. As such, detailed descriptions of the available surveys were confined to the SEAMAP, NMFS and VIMS survey.

5.3.1 SEAMAP

5.3.1.1 Sampling Intensity Samples were taken by trawl from the coastal zone of the South Atlantic Bight (SAB) between Cape Hatteras, North Carolina, and Cape Canaveral, Florida. Multi-legged cruises were conducted in spring (early April - mid-May), summer (mid-July - early August), and fall (October - mid-November). Stations were randomly selected from a pool of stations within each stratum. The number of stations sampled in each stratum was determined by optimal allocation. A total of 102 stations were sampled each season within twenty-four shallow water strata, representing an increase from 78 stations previously sampled in those strata by the trawl survey (1990-2000). Strata were delineated by the 4 m depth contour inshore and the 10 m depth contour offshore. In previous years, stations were sampled in deeper strata with station depths ranging from 10 to 19 m in order to gather data on the reproductive condition of commercial penaeid shrimp. Those strata were abandoned in 2001 in order to intensify sampling in the shallower depth-zone.

29

The R/V Lady Lisa, a 75-ft (23-m) wooden-hulled, double-rigged, St. Augustine shrimp trawler owned and operated by the South Carolina Department of Natural Resources (SCDNR), was used to tow paired 75-ft (22.9-m) mongoose-type Falcon trawl nets without TED’s. The body of the trawl was constructed of #15 twine with 1.875-in (47.6-mm) stretch mesh. The cod end of the net was constructed of #30 twine with 1.625-in (41.3-mm) stretch mesh and was protected by chafing gear of #84 twine with 4-in (10-cm) stretch “scallop” mesh. Trawls were towed for twenty minutes, excluding wire-out and haul-back time, exclusively during daylight hours (1 hour after sunrise to 1 hour before sunset). Contents of each net were sorted separately to species, and total biomass and number of individuals were recorded for all species of finfish, elasmobranchs, decapod and stomatopod crustaceans, cephalopods, sea turtles, xiphosurans, and cannonball jellies. Only total biomass was recorded for all other miscellaneous invertebrates (excluding cannonball jellies) and algae, which were treated as two separate taxonomic groups. Published characteristics of the fishing gear (sweep, headrope height) can be found in Stender and Barans (1994).

5.3.1.2 Biological Sampling In every collection, each of the priority species was weighed collectively and individuals were measured to the nearest centimeter. For large collections of the priority species, a random sub-sample consisting of thirty to fifty individuals was weighed and measured. Depending on the species, measurements were recorded as total length, fork length, or carapace width. Additional data were collected on individual specimens of penaeid shrimp (total length in mm, sex, female ovarian development, male spermatophore development, occurrence of mated females), blue crabs (carapace width in mm, individual weight, sex, presence and developmental stage of eggs), sharks (total and fork lengths in cm, individual weight, sex), horseshoe crabs (prosomal width and length in mm, individual weight, sex), and sea turtles (curved and straight lengths and widths in cm, individual weight, PIT and flipper tag numbers). Marine turtles were released in good condition according to NMFS permitting guidelines. Gonad and otolith specimens were also collected during seasonal cruises. A representative sample of specimens from each centimeter size range within each stratum were measured to the nearest mm (TL and SL), weighed to the nearest gram, and assigned a sex and maturity code (Wenner et al. 1998). Sagittal otoliths and a representative series of gonadal tissue were removed, preserved, and transported to the laboratory at MRRI, where samples were processed. Hydrographic data collected at each station included surface and bottom temperature and salinity measurements taken with a Seabird SBE-19 CTD profiler, sampling depth, and an estimate of wave height. Additionally, atmospheric data on air temperature, barometric pressure, precipitation, and wind speed and direction was also noted at each station.

30

5.3.1.3 Aging Methods

A detailed description of the aging methods used for Atlantic croaker from the SEAMAP collections is described in Wenner (2003).

5.3.1.4 Development of Estimates

Standardized estimates were determined using a delta-lognormal General Linear Model (Lo et al. 1992; Williams 2001). The proportion of positive tows was modeled using a binomial GLM and the positive tows using a lognormal GLM. Explanatory terms used in the model were year, season, and strata. Error estimates were obtained from a bootstrap procedure which re-samples residuals from the lognormal GLM model of the positive values and randomly draws values from the binomial distribution based on the observed and predicted positive data (Williams, 2001). Coast-wide estimates and regional estimates were developed for the SEAMAP index (see section 6.2. for rationale). Regional estimates consisted of a northern region and southern region, which split the survey into north of North Carolina-South Carolina border.

5.3.2 NMFS Northeast Trawl Survey

5.3.2.1 Sampling Intensity The NMFS Northeast Trawl Survey is the longest running continuous time series of research vessel sampling in the world and comprises of two seasonal surveys; Spring and Fall. The fall survey was initiated in 1963; the spring in 1968. These surveys cover the ocean environment from 5 to 200 fathoms deep, from Cape Hatteras, North Carolina to well beyond the Canadian border. About 300 half-hour trawl sets are made at sites randomly chosen prior to the beginning of each survey. The distribution of trawling locations is allocated according to a statistical method that divides the region into a number of smaller areas with similar depth characteristics. Detailed descriptions of the survey and annual survey reports can be obtained at: http://www.nefsc.noaa.gov/femad/ecosurvey/mainpage/survey.htm

5.3.2.2 Biases Recently it was found that the marks on the cable attaching scientific survey gear to the vessel Albatross IV were not at the 50 m length intervals they intended to indicate. The vessel crew used these marks to determine how much cable is deployed. The cable was most recently replaced in February 2000, and used in eight bottom trawl surveys, beginning with Winter 2000 and ending with Spring 2002, which may have an impact on estimates in these years, by affecting the “catchability”. Details of the problem can be found at: http://www.nefsc.noaa.gov/survey_gear/ .

31

5.3.2.3 Aging Methods Not Applicable.

5.3.2.4 Development of Estimates Estimates for Atlantic croaker were developed using the fall survey (Table 5.3.2.4.1). Standardized estimates were determined using a delta-lognormal General Linear Model (Lo et al. 1992; Williams 2001). The proportion of positive tows was modeled using a binomial GLM and the positive tows using a lognormal GLM. Explanatory terms used in the model were year, depth and latitude. Error estimates were obtained from a bootstrap procedure which re-samples residuals from the lognormal GLM model of the positive values and randomly draws values from the binomial distribution based on the observed and predicted positive data (Williams 2001).

5.3.3 Virginia Institute of Marine Sciences (VIMS) Trawl Survey. The VIMS trawls survey began in 1955. It a young of the year survey that samples “…. from the mouth of the Chesapeake Bay up to the freshwater interface at the fall line of the James, York, and Rappahannock Rivers”. For Atlantic croaker, the spring index is considered a more reliable measure of young of the year abundance and is representative of fish up to 100 mm (VIMS, Personal Communication). It is believed that Atlantic croaker are very sensitive to cold winter temperatures are susceptible to winter die-offs (VIMS, Personal Communication). As such, the spring index is considered a better indicator of young-of-the-year. Details of the survey can be found at http://www.fisheries.vims.edu/trawlseine/mainpage.htm

5.3.3.1 Sampling Intensity The VIMS index has a long time series of over 40 years. An annual sample size for the spring index ranged between 106 and 591 for the time period 1973-2002. The average annual number of samples was 295 samples/year. These samples include bay and river stations.

5.3.3.2 Biases Unknown

5.3.3.3 Biological Sampling Unknown

5.3.3.4 Aging Methods Not Applicable

5.3.3.5 Development of Estimates

Data were provided by VIMS. Estimates presented are those based on the geometric mean, and takes into account gear conversion factors (Table 5.3.3.5.1).

32

5.3.4 Length/Weight/ Catch-at-Age See section 6.2.2 for use of length-weight data. A catch at age matrix was not developed for this assessment.

5.3.5 Abundance Indices

No population abundance indices were available. See section 6.2.1 for indices used in the assessment and the assumptions made on the portion of the population they represented.

5.3.6 Biomass Indices

At present there are no biomass indices that represent the Atlantic croaker population.

5.3.7 Natural Mortality Estimates A key parameter in a stock assessment is the natural mortality (M). However, it is also a parameter that is difficult to ascertain for species that have been exploited. Estimates of M are usually obtained using life history analogies, which have been expressed in terms of equations or rules of thumb (e.g. Pauly (1980), Hoenig (1983), Gabriel et al. (1989)). In the early to mid 1990’s (Barbieri et al. 1994b) and in the previous assessment of Atlantic croaker (Lee et al. 2001), the maximum observed age based on available information was considered to be between 7-8 years. Lee et al. (2001) used an M=0.35 for their base model. However, recent data indicate a maximum observed age of 12 years from commercial landings in Virginia and North Carolina. Otoliths collected from Indian mounds ~ 1400-1700 indicate a maximum age of 15 years for Atlantic croaker during a period in which, presumably, the exploitation was much lower than in recent times. Mortality estimates using these different data sources indicate that the estimates using the archeological data were the lowest, with data from the mid-1990’s producing the highest estimates (Table 5.3.7.1). In general, adjusting for a fishing mortality rate of 0.1 for data collected in recent times, an estimated natural mortality rate range would be 0.15-0.4 and 0.2-.28 for the archeological data using Hoenig’s (1983) method or 3/maximum age rule of thumb. Estimates using Pauly’s (1980) and Alverson and Carney’s method (Quinn and Deriso 1999) produced higher natural mortality estimates between 0.2-0.7. In addition, the total mortality, Z necessary to result in the observed proportion at the oldest age for each year was estimated for the North Carolina and VIMS/ODU data. Z estimates between 1998-2002 ranged from 0.78-0.44 for the North Carolina data and 0.59-0.39 for the VIMS/ODU data set.

6.0 Methods

6.1 Model(s) For this analysis the primary data sources considered were the commercial and recreational landings, a selection of fishery independent juvenile to age 2 indices, the MRFSS index, together with age data from North Carolina (1996-2002) and Virginia (1998-2002).

33

Based on the available data, two model approaches were identified for a detailed evaluation. For this assessment we used a non-equilibrium surplus production model (ASPIC, Prager 1994) and age-structured surplus production model (Punt et al. 1995) to evaluate the population status of Atlantic croaker.

6.1.1. Surplus Production Model

The non-equilibrium surplus production model was implemented using ASPIC (Prager 1994). The foundation of the non-equilibrium surplus production can be described by the equation:

2ttt

t BK

rBFr

dt

dB

Where Bt = population biomass at time t; Ft= fishing mortality at time t; r = the stocks intrinsic rate of increase and K= maximum population size (carrying capacity). In the model, the fishing mortality rate for each fishery at time t is defined as:

jtjjt fqF

Where Fjt = Fishing mortality for fishery j at time t, qj = the catchability coefficient for fishery j, and fjt = effort for fishery j at time t. The data required for fitting the model are catch, effort, or yield for each time period, and can include population biomass indices for the time periods. The model estimates the initial biomass, r, K, and a catchability coefficient for each fishery that is included. Details of the objective function can be found in Prager (1994).

6.1.2. Age Structured Production Model.

The age-structured model we used is similar in structure to a forward-projection catch-at-age-model, with the exception that the some of the parameters are deterministic and based on available information.

The model uses a deterministic age-structured model to explain the population dynamics of a species where the population in successive years was linked using a Beverton and Holt stock recruitment relationship re-parameterized in terms of steepness. The major deterministic components in the model were parameters that characterized the growth, fecundity, and morphometrics of the species and; selectivity patterns for all of the fisheries and indices included in the model. To obtain a solution, the model minimizes the objective function by estimating a fully recruited fishing mortality rate of each year and fishery, catchability coefficients for the indices, virgin recruitment R0 and a set of annual recruitment deviations from the stock-recruit relationship.

The population abundance for the model is estimated using the following equations:

For the initial year:

34

10exp1

exp

91exp

0

)(

)(

,1,

)(,1,

:0,

,1

,1

,1

awhereNN

awhereNN

awhereSSBRN

ya

ya

ya

FM

FM

yaya

FMyaya

ratiovirgininitya

R0 represents the virgin recruitment estimated by the model; M = natural mortality and F = fishing mortality at age, a in year, y. SSBinit:virgin ratio is the ratio of the spawning stock biomass in the initial year to the virgin spawning biomass and was user defined as 0.75 in the base model.

For all other years:

10expexp

91exp

0exp)2.0()1(2.0

8.0

)(,

)(1,1,

)(1,1,

101

1,

1,1,1

1,1

awhereNNN

awhereNN

awherehSSBhSSBRSSBR

hSRN

yaya

ya

y

FMya

FMyaya

FMyaya

yFyo

yoya

R0 represents the virgin recruitment estimated by the model; h is the steepness parameter (defined as the proportion of virgin-stock recruitment production that occurs at 20% of the virgin spawning stock size); SSBy-1 represents the spawning stock biomass during the previous year and SSBRF=0 is the calculated spawning stock biomass per recruit under no fishing. y represents the recruitment deviation from the Beverton and Holt stock recruitment relationship in year, y; M = natural mortality and F = fishing mortality at age, a in year, y.

The spawning stock biomass for a given year was estimated by:

yaya

a

a yay WtMaturityNSSB ,,

10

0 ,5.0

Where Na,y= numbers at age in year y; Maturitya,y = proportion of mature fish at age a in year, y; Wta,y= weight at age in year y in Kg.

The spawning stock biomass per recruit under no fishing (SSBRF=0) was estimated as:

35

10exp1

exp

91exp

01

1

1

10

00

awhereSSBRSSBR

awhereSSBRSSBR

awhereSSBR

where

WtMaturitySSBRSSBR

M

M

aa

Maa

a

aa

a

a aF

The predicted total catch for a given year was estimated using the Barnov catch equation, summed across fleets and ages.

yaz

ya

yanfleet

fleet

a

a yay WtZ

FNCatch ya

,,

,

1

10

0 ,

^

)exp1( ,

Where, Na,y = estimated population numbers at age a in year y. Fa,y = fishing mortality at age, a in year, y. Za,y = total mortality at age, a in year, y. Wta,y= individual weight at age in year y in metric tons. The predicted estimates for the indices were:

weightinisindexwhereySelectivitWtNqI

numbersinisindexwhereySelectivitNqI

yaindex

yaindex

Zpartyra

a indexayayaindexy

Zpartyra

a indexayaindexy

)(10

0 ,,,

)(10

0 ,,

,

,

exp

exp

Where Na,y = estimated population numbers at age a in year y. Za,y = total mortality at age, a in year, y. Wta,y= individual weight at age in year y in Kg. Selectivitya,index= selectivity for age a in index. Partyr index= the proportion of the year that has passed at the mid-point of the survey’s measurement represented as a fraction of the year. qindex is the estimated catchability coefficient for the index. The Objective function in the model was constructed of components representing the differences between observed and predicted catches and indices, together with a component that accounted for the amount of recruitment deviation from the spawner-recruit relationship. The total likelihood was described by:

devrecindexcatchtotal LLLL

The likelihood components for the catch were defined as:

n

nfleet

fleetynyncatch landingsPredictedlandingsL

1

2,, ))log()(log(

36

Where landingsn,y =observed landings for fishery n in year y. Predicted landingsn,y =predicted landings for fishery n in year y. λn= user assigned weighting component for fleet n (where the weighting component could be configured to represent 1/22) The likelihood component for the indices were defined as:

n

nindex

indexynynindex IndexPredictedIndexObservedL

1

2,, ))log()(log(

Where Observed indexn,y =observed estimate for index n in year y. Predicted Indexn,y =predicted estimate for index n in year y. λn= user assigned weighting component for index n (where the weighting component could be configured to represent 1/22) The likelihood component for the constraint recruitment deviations from the Beverton and Holt spawner-recruit curve were defined as:

2002

1973

2)(y

yydevrec RecDevL

Where RecDevy= the estimated log recruitment deviations in year, y with a mean=0 and λ = a user assigned weight (where the weighting component could be configured to represent 1/22).

6.2 Model Calibration Examination of the two major coast-wide fishery independent indices revealed differing trends (Figure 6.2.1.) For the NMFS survey, a relatively flat trend with a recent spike was observed, while for the SEAMAP index, the general annual trend revealed relatively high values in the early 1990’s followed by a relatively flat trend in recent years. A more detailed examination of the two indices by strata revealed that the divergent patterns reflected the geographical regions where sampling occurred (Figure 6.2.2). At the southern range of the species (Florida to South Carolina) the catch trend revealed higher estimates between 1989-1992 followed by a relatively flat pattern in recent times. For the northern range of the species (North Carolina to New York) normalized estimates by strata appear to be higher in recent times (1998-2002) with relatively low estimates in the early part of the time series. The working group identified three possible alternatives to splitting the analysis into two regions: 1) splitting the region at Cape Hatteras NC, where all data north of Cape Hatteras, NC were included in a mid-Atlantic model and all data south of Cape Hatteras were included in a south-Atlantic model; 2) including North Carolina’s ocean landings with the northern range and North Carolina’s bay (sounds) landings with landings from South Carolina to east Florida; 3) Splitting the region at the North Carolina- South Carolina border. Based on a consensus of the stock assessment-working group, it was decided to evaluate the population status of Atlantic croaker using two independent models for the northern region of the species (North Carolina and North) and the southern range of the species (South Carolina to Florida). Examination of the MRFSS recreational index also revealed a north-south split (Section 5.2.4). While there are no genetic differences between the northern and southern range of the

37

Atlantic croaker population, the dynamics of the regions may be different. At the southern range of species, the fishery is predominantly a small recreational fishery. In the south, the commercial fishery, only occurs in Florida, and was affected by the constitutional amendment than banned the use entangling fishing gear in state waters in the mid 1990’s. For years prior to 1981, when recreational landings were unavailable, total annual landings were estimated for each state using an adjustment factor derived from the ratios of annual commercial landings to annual total landings from 1981-2001. The state-specific adjustment factors were derived in one of two ways. For North Carolina and Florida – East Coast, the ratio of commercial to total landings appeared to be relatively stable (or at least did not exhibit marked trends) over time. For these states, the adjustment factor was the average of the commercial fraction of the total annual landings. In North Carolina for example, the average commercial fraction of total annual landings was 0.95 (i.e. commercial landings accounted for 95% of total annual landings, on average). To hind-cast total annual landings for North Carolina prior to 1981, the annual commercial landings was divided by 0.95. For New Jersey, Maryland, and Virginia, a temporal pattern was present that suggested when commercial landings were relatively large the commercial fraction of total annual landings was larger than in years when commercial landings were relatively small. For these states, two separate adjustment factors were calculated as (1) the average of the commercial fraction of the total annual landings in high commercial landings years (1984-1989,1993-2001), and (2) the average of the commercial fraction of the total annual landings in low commercial landings years (1981-1983,1990-1992). For these three states, total annual landings were calculated by first classifying years prior to 1981 as high or low based on their annual commercial landings and then dividing the year’s annual commercial landings by the respective high or low adjustment factor. Commercial landings data were unavailable for 2002. For each state, total annual landings were estimated as annual recreational landings divided by (1 – commercial adjustment factor). For New Jersey, Maryland, and Virginia, 2002 was classified as a high year based on the relatively high commercial landings in recent years (1997-2001). In this analysis, commercial discards have not been taken into consideration for the reasons outlined in Section 5.1.3. Estimates of recreational discards by weight were based on assuming that discarded fish were likely to be representative of fish equal to or lower than the 10th percentile of the size distribution of those fish measured (see section 5.2.1.6). As part of the preliminary analysis, a series of model runs were carried out. These comprised of two types. The first, were a set of core models for each of the regions (Mid-Atlantic and South-Atlantic). The core group of model runs consisted of a base run and eleven sensitivity runs for each region based on a set of natural mortality and steepness estimates. The base runs for each of the regional models, used a natural mortality rate of 0.3 and a steepness values associated with the 50th percentile of the posterior probability distribution for steepness (Figures 6.2.3 – 6.2.4). The second set of models evaluated the model sensitivity to configuration and input data. Some of the factors evaluated included: use of a two-fleet coast-wide model with the mid-Atlantic indices; the effects of increased weighting of the likelihood terms for the indices; effects of assuming alternate biological characteristics; use of alternate indices; and sensitivity to miss-

38

specification of selectivity in the commercial fishery. A summary document describing the results of these preliminary runs is available on the FTP site. Based on the preliminary analyses, the Atlantic croaker technical committee (ACTC) concluded that while, there was no genetic evidence to suggest two separate stocks, the population trends seen in the two regions differed, and the best way to capture those differences were through two separate models for the northern (mid-Atlantic model) and the southern regions of the stock (south-Atlantic model; Figure 6.2.5). The south Atlantic model represents the stock at its southern boundary. Personal observations of some members of the ACTC indicate that larger and older Atlantic croaker were rare in the South Atlantic. The lack of larger older fish could be the result of higher mortality rates or movement of older fish out of the region. The trends and estimates from a coast-wide model done in the preliminary analysis revealed that the estimates from the coast-wide model were almost identical to those of the mid-Atlantic model. Based on the preliminary analysis, the ACTC also identified four major factors that the model was sensitive to. These were: the natural mortality estimate, steepness parameter, the ratio of spawning stock biomass in 1973 relative to virgin conditions, selectivity estimates of Age 0-1 Atlantic croaker in the commercial fishery, and age 0 fish in the recreational fishery. The sensitivity of the model to these factors were examined through a series of Monte Carlo trials over a range of estimates (Table 6.2.1).

6.2.1 Tuning Indices For this analysis eight fishery independent surveys and two fishery dependent indices were available. An inspection of the fishery independent indices revealed that they were primarily targeting juveniles to Age 1. The eight fishery independent indices were the NMFS fall trawl survey, SEAMAP trawl survey, VIMS trawl survey, North Carolina DMF juvenile estuarine and sound surveys, Maryland DNR juvenile index, and the Florida FWC fishery independent trawl and seine surveys. The fishery dependent surveys were the MRFSS total catch index and the North Carolina DMF commercial CPUE index. Table 6.2.1.1 and Figures 6.2.1.1-6.2.1.3 summarize the annual estimates for the available fishery independent and dependent indices. For the SEAMAP index, estimates for the North, South, and combined estimates are included. For this analysis, our choice of indices was based on those that had the best spatial representation of the region over the time period. As such, for the mid-Atlantic region, the core indices considered were the NMFS trawl survey, SEAMAP survey, and MRFSS index. The VIMS spring juvenile index was included in one of the runs in the preliminary analysis, as it covered the entire time period 1973-2002. For the south-Atlantic region we used the SEAMAP and MRFSS indices. 6.2.2. Input Parameters and Specifications The input parameters required for implementing the model are:

1) Selectivity patterns for each of the indices used and the commercial and recreational fisheries.

39

2) Biological characteristics of the species described by the von Bertalanffy parameters (L∞, k, t0), length-weight relationship and maturity schedule.

3) Estimates of natural mortality, steepness, and the ratio of the initial years spawning stock biomass to the virgin spawning stock biomass.

4) Weightings for the likelihood components. 5) Bounds for parameters estimated by the model

Selectivity Estimated selectivity patterns for the fisheries and indices were initially determined by examining the available information on size and age range of the respective fisheries/indices, together with input from the members of the stock assessment-working group. The selectivity pattern for each fishery and index was specified by assigning a probability of capture for each age class in the model. For the commercial fishery, selectivity patterns were based on an examination of the size and age distribution of the fishery in the mid-Atlantic. Over the time series, the majority of commercial landings ranged between 220 –400 mm TL (Figure 6.2.2.1) and based on the North Carolina growth parameters, this size range would most likely have compromised of fish between Age 1 to 5 years (Figure 6.2.2.2.). An overlay of the 200-400 mm size range on the North Carolina age data is shown in Figure 6.2.2.3. In the early model runs a semi-observed catch-at-age matrix was used to iteratively tune age-specific selectivity estimates for the mid-Atlantic commercial fishery on a gross scale. Sufficient age data were available from Virginia’s commercial fishery to construct annual age-length keys for years 1998-2001. The observed catch-at-age matrix for Virginia’s commercial fishery, which included ages 1-10, was scaled to represent the mid-Atlantic region by multiplying each cell in the matrix by a year-specific expansion factor. The annual expansion factor was the ratio of the mid-Atlantic region annual commercial landings (NJ + MD + VA + NC) to Virginia’s annual commercial landings. The resulting semi-observed catch-at-age matrix for the mid-Atlantic region was used to calculate catch-at-age residuals ([semi-observed catch-at-age matrix] – [predicted catch-at-age matrix]) from initial model runs. Consistent patterns in the residuals indicated that the selectivity estimates were high for ages 1-2 (negative residuals, predicted catches too large), reasonable for ages 3-4 (residuals near 0), and low for ages 5-10 (positive residuals, predicted catches too small). Selectivity estimates were adjusted, and residuals were re-examined graphically after subsequent model runs. Formal optimization of selectivity parameters was not performed given the order-of-magnitude accuracy level of the semi-observed catch-at-age matrix. Based on the evaluations, a flat-topped selectivity pattern was used for the commercial fishery in all runs. Given, the lack of information of size for the south Atlantic fishery, the mid-Atlantic selectivity pattern was also used for the south Atlantic model. Based on the preliminary analyses, the ACTC concluded that there was much uncertainty on the selectivity estimates for Age 0 – 1 and modified the selectivity pattern accordingly. The available commercial data did not fully capture the size and age ranges of Atlantic croaker captured by the trawl fisheries. The ACTC concluded that as a base case for both models, a selectivity of 0.1 for age 0, 0.55 for Age 1 and 1.0 for all other ages was appropriate (Table 6.2.2.1).

40

The size distribution of Atlantic croaker caught by the recreational fishery was similar between for the mid-Atlantic and south Atlantic. However, larger fish between 295- 395 mm TL were less well represented in the south Atlantic fishery. However, the differences did not warrant a separate selectivity patterns for the regions. In general fish between 180 and 380 mm TL are well represent in the fishery (Figure 6.2.2.4). When the size estimates were converted to ages, fish age 1-8 are likely to be well represented in the fishery (Figures 6.2.2.5 and 6.2.2.6). For the recreational fishery, the ACTC modified the initial estimates of selectivity to include a selectivity estimate of 0.05 for age 0 fish and 1.0 for all other ages. For the NMFS trawl and SEAMAP trawl surveys the majority of Atlantic croaker caught was between 120-240 mm TL and 110 –230 mm TL respectively (Figure 6.2.2.7; Figure 6.2.2.10). It appears that both these indices are good indicators of Age 0-1 Atlantic croaker and the selectivity patterns chosen reflect this (Figures 6.2.2.8, 6.2.2.9, 6.2.2.11, 6.2.2.12; Table 6.2.2.1). For 2001 and 2002 age data from the SEAMAP survey also indicate that the majority of Atlantic croaker were age 0-1. Biological characteristics Age information on Atlantic croaker was available form five data sources. Based on an examination of the growth curves from those data sets, and estimates in the literature, the stock assessment-working group concluded the most appropriate growth model to assign an estimated length at age was that based on the North Carolina DMF data set on pooled sexes. As such, length at age was estimated using:

9572.1,2415.0,6.434

)exp1(

0

))(( 0

tkmmLwhere

LAgeatLength tAgek

Weight at age (kg) was estimated using:

13.3,1049.5

)(9

bawhere

AgeatLengthaAgeatWeight b

Based on Barbieri et al. (1994a) we assigned Atlantic croaker a maturity schedule where Age 0 fish were considered immature (0% maturity), by Age 1, 90% were mature and from Age 2 onwards, 100% were considered mature. Natural mortality and steepness Given the large range of natural mortality estimates derived using the traditional methods of approximation (~ 0.15-0.6), a range of natural mortality rates was chosen for the preliminary analyses (0.2 to 0.4). For the base models presented in the assessment, M=0.30.

41

Steepness, h is a measure of recruitment when the spawning stock biomass is reduced to 20% of the stock when no fishing is present. Steepness is an indicator of the ability of a fish stock to withstand high fishing mortality rates; high steepness values indicate a resilient species, as recruitment is high even when the spawning stock biomass is reduced to low levels. The choice of steepness values is an important factor in the model implementation, and is important in assessing the additional mortality a population can sustain over the long term (Myers et al. 2002). One solution to using appropriate estimates of steepness is to use a Bayesian approach in the model formulation. However, one of the most important considerations in using the Bayesian approach is having an informative prior distribution. Myers et al. (2002), have addressed this for Atlantic croaker by developing a series of prior distributions for the species using an empirical Bayesian approach. For this assessment we initially developed a posterior distribution for steepness for each of the regions using the Myers et al. (2002) prior distribution for Atlantic croaker (derived from their covariate analysis). The prior distribution was described using a beta distribution where 1=4.7728 and 2=2.2201(Figure 6.2.2.13). In the preliminary analyses, steepness values at various percentiles of the posterior distribution were used for the model runs. An examination of the steepness estimates from the posterior probability distribution indicated a median estimate of 0.5 for both regional models. On further inspection, the ACTC concluded that the data provided little to no information for steepness. As such, it was decided to use the modal estimate from Myers et al (2002) prior as the base value of steepness (0.76). Weighting of the likelihood components Weighting of the likelihood components can have an important effect on the outcome on the estimates of the parameters. The main likelihood components in the model are based on a lognormal distribution. For all analyses we gave all components, except the fishery independent indices, a weighting (λ) of 1. This was equivalent to assigning the residuals in each data set a coefficient of variation (C.V.) of 0.8. (Where C.V.= √(exp(σ2)-1) as λ is equivalent to 1/ 2 σ2. For the fishery independent indices the weightings were increased to 2.0 (equivalent to a C.V. of 0.5). The ACTC felt that the residuals of the fishery independent indices were likely to be associated with less variability than the other terms in the objective function. Bounds for parameters estimated by the model For models implemented in Excel, the fishing mortality rate per fleet was bound between 0 and 1.5. The parameter bounds for all models implemented in AD model Builder were similar, and are summarized in Table 6.2.2.2.

7.0 Outputs/Results

The preliminary runs for the surplus production model revealed it was unstable. The model was highly sensitive to the model inputs, with the estimates of MSY and B1/K approaching the bounds. As such, the stock assessment group decided not to move further with this model. In addition, there were some other concerns with this model, as the available indices did not

42

represent the population biomass, as they predominantly represented juveniles to Age 1. Results presented in this analysis were based on the model implemented in AD model builder and Excel. The results of the EXCEL based model were similar to those produced using the AD model builder version (See Appendix A for a comparison of similarly configured models).

7.1 Goodness of Fit of Model Used

The goodness of fit of a statistical model is judged by how well the predicted estimates match the observed estimates. The residuals for recreational and commercial landings indicated a good fit for the mid-Atlantic and south Atlantic models (Figures 7.1.1-7.1.4; Table 7.1.1). However, for the south Atlantic model, the high recreational landings in 1984 and 1986 were poorly estimated. Residuals were associated with low standard deviations; the mean and standard deviation of the residuals for each of the fleets and their standardized-residuals for the mid-Atlantic and south Atlantic base models are presented in Table 7.1.1. Estimates of the standard deviation of the residuals and the standardized residuals indicated that for the mid-Atlantic, the MRFSS index was associated with best fit (residual mean=0 and standard deviation=0.52 for base model) and appears to be an important index influencing the model. The residuals standard deviations for the SEAMAP index were in a similar range to the MRFSS index (mean=0 and standard deviation =0.65). The NMFS trawl survey was associated with the greatest variability (Figures 7.1.5-7.1.7; Table 7.1.2). For the south-Atlantic model, the model appeared to fit the data reasonably well, except for the high points associated with both indices (Figures 7.1.8-7.1.9; Table 7.1.2). Based on the standard deviation of the residuals, the MRFSS and SEAMAP index appeared to have influenced the model in approximately equal proportions. In general, few data points exceeded an absolute value of 2.0 for the standardized residuals in either the mid-Atlantic or south-Atlantic base models.

7.2 Parameter Estimates

The model estimates a total of 93 parameters for the mid-Atlantic model and 92 parameters for the south-Atlantic model. The estimated parameters include an annual fully selected fishing mortality rate for each fishery, an annual recruitment deviation from the stock-recruitment relationship, the number of virgin recruits (R0) and a catchability coefficient for each of the indices.

7.2.1 Exploitation Rates (should include both F and u) Unless otherwise noted, fishing mortality rates referred to in the document are the combined fully selected fishing mortality rate for the commercial and recreational fishery. Exploitation rates were estimated using the fully selected fishing mortality rate for the commercial and recreational fishery combined. For the mid-Atlantic region, a cyclical trend in fishing mortality rates is apparent, with the highest fishing mortality rates occurring in the mid 1970’s, followed by a cyclical peak in the mid 1980’s and again between 1997 and 1998 (Figure 7.2.1.1). However, the most recent peak in

43

fishing mortality appears to be lower than those observed in the past, ranging around 0.5 per year (Table 7.2.1.1). For the mid-Atlantic fishery, the recreational fishery accounts for a relatively small proportion of the total fishing mortality (Table 7.2.1.1). In the south-Atlantic region, a cyclical pattern to fishing mortality rates was also observed, with the highest peak between 1986-87 (Figure 7.2.1.2). More recently, fishing mortality rates have peaked in 2000-2001 (Table 7.2.1.1). For the south-Atlantic fishery, the recreational fishery accounted for the largest proportion of the total fishing mortality during the time series. In the model, fishing mortality estimates were limited to a maximum of 1.5. Estimates of recreational fishing mortality for the south-Atlantic model frequently hit this upper bound (Table 7.2.1.1) Exploitation rates for the base mid-Atlantic and south-Atlantic models are presented in Tables 7.2.1.2. In general, the trends in exploitation rates mirror those of the fishing mortality rates. For the base mid-Atlantic model, the exploitation rate between 1999 and 2002 has been at around 0.25 (Table 7.2.1.2). For the base south-Atlantic model, exploitation rates between 1999 and 2002 have ranged between 0.33 and 0.59.

7.2.2 Abundance Estimates Estimates of abundance in numbers for the base runs for the mid-Atlantic and south –Atlantic models are presented in Tables 7.2.2.1. For the mid-Atlantic region, the trend in population abundance indicates a step-wise increase in population size reaching a peak in 1998. For the base mid-Atlantic run, population estimates from 1999 to 2002 have ranged around 500 million fish (Table 7.2.2.1). For the south-Atlantic model, the population trend is an inverse to that observed for the mid-Atlantic model. Estimated population sizes were high in mid 1980’s and during the recent part of the time series have been relatively stable and low. For the base south-Atlantic run, population estimates between 1999 and 2002 have range between 3-4 million fish. For the mid-Atlantic model, spawning stock biomass (expressed as the proportion of mature females) shows a sharp decline from the early 1970’s up to the early 1980’s. From 1981 to 1991, estimates of spawning stock biomass shows relatively flat trend up to the early 1990’s (Figure 7.2.2.1). From the early 1990’s spawning stock biomass has increased sharply and since 1999 has been relatively stable and high at around 30,000 metric tons (Table 7.2.2.1). For the south-Atlantic model, spawning stock biomass was highest in the early part of the time series, decreased in a stepwise pattern and remained relatively stable since 1995 (Figure 7.2.2.2). For the base south-Atlantic model spawning stock biomass estimates between 1999-2002 have ranged between 130-170 metric tons (Table 7.2.2.1). The recruitment trend for the base mid-Atlantic model reveals a relatively flat recruitment period up to the early 1980’s. From the early 1980’s onwards, periodic spikes in recruitment in 1983, 1991, 1994, and 1998 occurred (Figure 7.2.2.1). Between 1999 and 2002 the estimated number Age 0 recruits for the mid-Atlantic base model has ranged between 390 and 526 million fish (Table 7.2.2.1). For the base south-Atlantic models the recruitment trend also reveals a cyclical pattern over the time series (Figure 7.2.2.2). Between 1999 and 2002 the number of Age 0 recruits for the base south-Atlantic model ranged between 1.3 and 2.4 million fish.

44

7.2.3 Precision of Parameter Estimates For models run using AD model builder, estimates of standard deviation are based on the delta method, which approximate the variance estimates. Variance estimates using the delta method are biased to the lower range of the spectrum when additional constraints are imposed on the model (ASMFC, in preparation). As such, a Monte-Carlo re-sampling scheme was used to examine the uncertainty surrounding the parameter estimates and described more completely in Section 7.4. 7.3 Projection Estimates No stock projections were carried out . 7.4 Sensitivity Analyses To evaluate the sensitivity of the mid-Atlantic model to the deterministic inputs (see section 6.2), the ACTC identified a subjective weighting for each of the sensitivity inputs. These weightings were used to create a probability distribution for all parameters, except steepness (Figure 7.4.1). For steepness, the prior distribution developed by Myers et al (2002) was used (Figure 6.2.2.13). Using the probability distributions, 1,299 runs were carried out using the re-sampling procedure and are summarized in Table 7.4.1. Commercial fishing mortality rates show a broad range of estimates over much of the time series (Figure 7.4.2 a). The effects of constraining the fishing mortality estimate to a maximum of 1.5 per year are clearly evident on some model permutations for the period 1978-1980. From 1997 onwards a declining trend in commercial fishing mortality rates is evident for all permutations examined. Fishing mortality rates in the recent part of the time series also appear to have less variability than estimates from the early years. Recreational fishing mortality rates per year indicate a close correspondence among the different trials (Figure 7.4.2. b). Spawning stock biomass estimates suggest a relatively broad range of estimates during certain periods (Figure 7.4.3). Examination of the trials indicates that the low SSB estimates were associated with high steepness values, (0.8-0.9) and the high estimates were associated with low steepness estimates (0.2- 0.4). Trends in Age 0 recruits among the trials suggest that recruitment patterns were consistent over the majority of trials (Figure 7.4.4). Low recruitment estimates were associated with high steepness values and high recruitment estimates were associated with low steepness values. 7.5 Retrospective Analyses To date, a traditional retrospective analysis, where the last few years in the data series were sequentially deleted and its effect evaluated, has not been done.

45

8.0 Biological Reference Points

Currently, there are no established biological reference points for Atlantic croaker. Based on the current model, the ACTC concluded that the most appropriate reference points would be those based on MSY criteria. The ACTC noted that, as more data become available and the assessment model evolves, alternate reference points should be considered. The ACTC also discussed the need for different reference points for the northern (mid-Atlantic model) and southern (south-Atlantic model) regions of the stock. However, the ACTC had concerns on recommending and evaluating reference points for the south-Atlantic model at this time. Personal observations of some members of the ACTC indicate that larger and older Atlantic croaker were rare in the South Atlantic. The lack of larger older fish could be the result of higher mortality rates or movement of older fish out of the region. Given the lack of information on movement rates of Atlantic croaker between the two regions, estimates of Fmsy and SSBmsy for the south Atlantic may be incorrect. 8.1 Overfishing Definition Restrepo et al. (1998) describe a set of biological reference points or benchmarks that are based on fishing mortality (maximum fishing mortality threshold) and spawning stock biomass (minimum stock size threshold) that relate to implementing National Standard 1 of the Mangnuson-Stevens Fishery Conservation and Management Act. The national standards guidelines identifies thresholds that are necessary to maintain a stock within safe levels and are used to determine if a stock is being overfished or is in a overfished state. Currently, there are no established definitions of over fishing for Atlantic croaker. As such, the ACTC believe that adoption of the default criteria suggested by Restrepo et al (1998) would be appropriate for the mid-Atlantic region.

1) Fishing mortality threshold, Fmsy 2) Fishing mortality target, 0.75 Fmsy 3) Biomass threshold, 0.5 SSBmsy 4) Biomass target, (1-M) SSBmsy =0.7 SSBmsy

Examination of the phase plots of the ratio of F2001/ Fmsy with SSB2001/SSBmsy and those based on average estimates between 1999-2001 from the Monte-Carlo simulations suggest that SSB/ SSBmsy ratios for a large proportion of the runs were greater than 1.0 (Figures 8.1.1-8.1.2;Table 8.1.1). F / Fmsy ratios for 50% of the runs were less than 1.04 (Table 8.1.1). Estimates from the base model were close to the median estimates from the Monte-Carlo trials. For the base mid-Atlantic model, fishing mortality rates in recent years have been close to Fmsy levels and above the proposed threshold level (Figure 8.1.3). Annual spawning stock biomass estimates from the base mid-Atlantic model in recent years has been above the proposed target and threshold levels (Figure 8.1.4).

46

8.2 Stock Recruitment Analysis As part of the model configuration, a Beverton and Holt stock recruitment relationship re-parameterized in terms of steepness is included. Estimates of the Virgin recruitment for the base mid-Atlantic and south-Atlantic models were 114 and 5.8 million fish respectively. The stock recruitment curves for the base mid-Atlantic and south-Atlantic model are presented in Figures 8.2.1 and 8.2.2. For the base mid-Atlantic model a wide scatter between recruits and spawning stock is evident; whereas for the south-Atlantic model the scatter implies a downward trend in recruitment over the time series. The limitations of using two independent models to estimate the stock of Atlantic croaker over its range should be noted when evaluating these plots. 8.3 Yield and SSB per Recruit Yield per recruit and SSB per recruit were estimated for the base mid-Atlantic and south Atlantic models. The selectivity pattern used was based on the average catch weighted selectivity over the last three years. In general, for both base models, the yield per recruit curve is flat (Figures 8.3.1 and 8.3.2). For the base mid-Atlantic model the average SPR over the last four years was 35 %, while for the base south-Atlantic model, the average SPR was13 %. 8.4 Stock Production Model The stock production model was found to be unstable, and its use in this assessment was discontinued.

9.0 Recommendations and Findings

The mid-Atlantic model, which is the core of the population, indicates fishing mortality rates were high in the mid 1970’s, abruptly declined, and has shown a cyclical trend in the mid 1990’s, and appear to have stabilized. A preliminary catch curve analysis using the North Carolina and VMRC/ODU age data suggest that total mortality rates from 1998 to 2002 have declined from around 0.6-0.8 in 1998 to 0.31-0.4 in 2002. Using an M=0.3, these estimates of total mortality compare favorably to the base mid-Atlantic model (Full-F=0.5 in 1998 and 0.26 in 2002). The commercial age data from recent years also shows an increasing age distribution, with fish of 12 years being observed in the commercial landings. The mid-Atlantic model is primarily driven by MRFSS and SEAMAP index, which describe an increasing abundance trend in recent years. In addition, the NEFC trawl index also lends support to increasing trends of juveniles. Anecdotal evidence from the mid-Atlantic range of the population in Delaware, suggests an abundance of Atlantic croaker in the region (D. Kahn, personal communication). The population has benefited from good recruitment in recent years, which may also be tied to the regulatory changes that have affected some of the fisheries that indirectly target Atlantic croaker (see Section 3.2). The southern region of the stock appears to have a different set of dynamics than that of the northern range. It is evident that the recent increases in recruitment seen in the mid-Atlantic have not been observed in the south. While the results of the model suggest high exploitation rates in the south, the migratory nature of the stock is not addressed in either of the models. Further, the role of the environment on the region is also poorly understood. It has been suggested that the

47

recent and prolonged drought may have had some impact on the local population. Furthermore, the MRFSS estimates for Atlantic croaker in the South-Atlantic are also relatively imprecise, with landings being associated with high PSE values (Table 5.2.2.3). More effort needs to be spent on evaluating the south-Atlantic model, before appropriate management benchmarks are developed for the region. In this assessment we examined the population in two independent models. Linking the two models into one, with the incorporation of additional data on movement patterns should be goal for future assessments. Treating the population as one by combining the landings together masks the dynamics of the southern range of the species. 9.1 Evaluation of current status based on biological reference points Based on the proposed reference points for the mid-Atlantic model, fishing mortality rates in recent years have been close to Fmsy levels and above the proposed threshold level (Figure 8.1.3). Annual spawning stock biomass estimates for the mid-Atlantic recent years has been above the proposed target and threshold levels (Figure 8.1.4). 9.2 Research Recommendations

The technical committee provided a prioritized listing of research recommendations, as shown below: 1. Need for more movement data from the south region, including tagging information from Cape Fear south. Examine otolith microchemistry data available and continue research in this area. 2. Need for bycatch and discard estimates from the commercial and recreational fisheries (i.e. shrimp fishery). Characterization of scrap fishery.

3. Standardize ageing procedures for Atlantic croaker and standardize current age data sets. Need for Coast wide collection of bio-profile information and add standardized protocols for those data.

4. Produce a general fishery independent index using state survey information. Develop a coast wide and or regional CPUE index.

5. Investigate including climatic factors in the model.

6. Need for an updated maturity schedule.

7. Examine socio-economic aspects of the fishery.

48

10.0 Literature Cited

Atlantic States Marine Fisheries Commission (ASMFC). 1987. Fishery Management Plan for Atlantic Croaker. Fisheries Management Report No. 10, Oct 1987. ASMFC. 1993. Proceedings of a Workshop on Spot (Leiostomus xanthurus) and Atlantic Croaker (Micropogonias undulates), L. Kline and H. Speir, eds. ASMFC, Washington DC. 160pp. ASMFC. 2002. 2002 Review of the Fishery Management Plan for Atlantic Croaker, Washington DC. ASMFC. In Preparation. Atlantic menhaden 2003 stock assessment report. Atlantic States Marine Fisheries Commission. Barbieri, L. R. 1993. Life history, population dynamics and yield-per-recruit modeling of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay area PhD Dissertation College of William and Mary, Williamsburg, VA. 140 p. Barbieri, L.R., M.E. Chittenden Jr., and S.K. Lowerre-Barbieri. 1994a. Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropogonias undulates, in the Chesapeake Bay and adjacent coastal waters. Fishery Bulletin 92: 671-685. Barbieri, L.R., M.E. Chittenden Jr., and C.M. Jones. 1994b. Age, growth and mortality of Atlantic croaker, Micropogonias undulates, in the Chesapeake bay region, with a discussion of apparent geographic changes in population dynamics. Fishery Bulletin 92:1-12. Barger, L.E. 1985. Age and growth of Atlantic croakers in the Northern Gulf of Mexico, based on otolith sections. Transactions of the American Fisheries Society 114: 847-850. Bobko, S.J., C.M. Jones and E.M. Robillard. 2003. Results of 2001 Virginia-Chesapeake Bay Finfish Ageing. VMRC/ODU Age and Growth Lab. Old Dominion University, Norfolk, VA. 67 pp. Chittenden Jr., M.E., L.R. Barbieri, and C.M Jones. 1994. Development of Age Determination Methods, Life History/Population Dynamics Information, and Yield-per-Recruit Simulation Modeling to Evaluate the Potential for Growth and Recruitment Overfishing of Atlantic Croaker, Micropogonias undulates, in the Chesapeake Bay. Cowan, J.H., Jr., R.L. Shipp, H.K. Bailey IV, and D.W. Haywick. 1995. Procedure for rapid processing of large otoliths. Trans. Am. Fish. Soc. 124:280-282. Diamond, S.L., L.B. Crowder, and L. Cowell. 1999. Catch and Bycatch: The qualitative effects of fisheries on population vital rates of Atlantic croaker. Transactions of the American Fisheries Society 128: 1085-1105.

49

Gabriel, W.L. M.P. Sissenwine and W.J. Overholtz. 1989. Analysis of Spawning Stock Biomass per recruit: An example for Georges bank haddock. North American Journal of Fisheries Management 9:383-391 Hales Jr., H.L. and E.J. Reitz. 1992. Historical changes in age and growth of Atlantic croaker, Micropogonias undulatus (Perciformes: Sciaenidae). Journal of Archaeological Science:19:73-99 Hoenig, J.M. 1983. Estimating total mortality rate from longevity data. Ph.D. Dissertation. University of Rhode Island Krebs, C.J. 1989. Ecological Methodology. Harper Collins Publishers. 654 pp. Lankford, T.E., Jr., T.E. Targett and P.M. Gaffney. 1999. Mitochondrial DNA analysis of population structure in the Atlantic croaker, Micropogonias undulatus (Perciformes: Sciaenidae). Fishery Bulletin 97:884-890 Lankford, T.E. Jr., and T.E. Targett. 2001. Low-temperature tolerance of age-1 Atlantic croakers: Recruitment implications for US Mid-Atlantic estuaries. Transactions of the American Fisheries Society 130: 236-249 Lee, Laura M., J.E. Hightower and P.S. Rand. 2001. Population dynamics of Atlantic croaker occurring along the U.S. east coast, 1981-1998. Lo, N.C.H., D. Jacobson, J.I. Squire. 1992. Indices of relative abundance from fish spotter data based on delta-lognormal methods. Can. J. fish. Aquat. Sci. 49:2515-2526. MRFSS 1999. MRFSS user’s manual. A guide to use of the National Marine Fisheries Service Marine Recreational Fisheries Statistics Survey Database. Atlantic Marine Fisheries Commission in cooperation with the National marine Fisheries Service and U.S. Fish and Wildlife Service. March 1999. Myers, R.A., N.J. Barrowman, R. Hilborn, D.G. Kehler. 2002. Inferring Bayesian priors with limited directed data: Applications to risk analysis. North American Journal of Fisheries Management 22:351-364. North Carolina Division of Marine Fisheries (NCDMF). 2001. Assessment of North Carolina Commercial Finfisheries, 1997-2000. Final Performance report for Award Number NA 76 FI 0286, 1-3. North Carolina Department of Environment and Natural Resources. North Carolina Division of Marine Fisheries.. North Carolina Division of Marine Fisheries (NCDMF). 2002. Survey of population parameters of marine recreational fishes in North Carolina. Annual Progress Report. Segment 11. North Carolina Department of Environment and Natural Resources. North Carolina Division of Marine Fisheries. 36 pp. Pauly, D. 1980. On the interrelationships between natural mortality, growth parameters, and

50

mean environmental temperature in 175 fish stocks. J. Cons.Int.Explor.Mer 39:175-92.

Prager, M.H. 1994. A suite of extensions to a non-equilibrium surplus-production model. Fishery Bulletin 92:374-389. Punt. A.E. , Butterworth, D.S. and J. Penny. 1995. Stock assessment and risk analysis for the South Atlantic population of Albacore Thunnus alalunga using an age-structured production model. South African Journal of Marine Science 16: 287-310. Reststrepo, V.L., G.G. Thompson, P.M. Mace, W.L. Gabriel, L.L. Low, A.D. MacCall, R.D. Methot, J.E. Powers, B.L. Taylor, P.R. Wade, J.F. Witzig. 1998. Technical guidance on the use of precautionary approaches to implementing National Standard 1 of the Magnuson-Stevens Fishery Conservation and Management Act. U.S. Dept. of Commerce. NOAA Technical Memorandum, NMFS-F/SPO-31. Ross, S.W. 1988. Age, growth and mortality of the Atlantic croaker. Transactions of the American Fisheries Society 117: 461- Stender, B. and C.A. Barans. 1994. Comparisons of the catch from tongue and 2-seam shrimp nets off South Carolina. North American Journal of Fish. Mgmt. 14: 178-195.

Quinn, T.J. and R.B. Deriso. 1999. Quantitative fish dynamics. Oxford University Press. Wenner, C.A., C. Walton, M. Brouwer and E.L. Wenner. 1998. Age, growth and reproductive biology of selected species taken incidentally to the penaeid shrimp fisheries along the southeastern coast of the U.S. Final Report for MARFIN Grant Number NA57FF0297. Wenner, C.A. 2003. Atlantic croaker, Micropogonias undulatus, along the middle Atlantic and southeast coast of the United States. A research proposal submitted to MARFIN. South Carolina Department of Natural Resources 21 pp. White, M.L., M.E. Chittenden Jr. 1977. Age determination, reproduction, and population dynamics of the Atlantic croaker, Micropogonias undulatus. Fishery Bulletin 75: 109-122. Williams, E.H. 2001. Assessment of cobia, Rachycentron canadum, in the waters of the U.S. Gulf of Mexico. NOAA Technical Memorandum NMFS-SEFC-469.

51

11.0 Tables

Table 122.1.1. Summary of available age data for Atlantic croaker ( only samples based on otolith readings were considered).

Source NC DMF VAMRC/ODUVIMS SEAMAP MARYLAND TOTAL

Type FI/FD FD FI/FD FI FD

1989 96 96

1990 32 32

1991-1995 0

1996 836 836

1997 428 428

1998 1,071 221 2,100 3,392

1999 671 317 2,260 180 3,428

2000 815 311 499 145 1,770

2001 793 364 797 38 1,992

2002 605 360 548 1,513

NC DMF -North Carolina Department of Marine Fisheries

VAMRC/ODU - Virginia Marine Resources Commission/ Old Dominion University

VIMS - Virginia Institute of Marine Science

SEAMAP- SEAMAP/South Carolina Department of Natural Resources

MARYLAND - Maryland DNR/ South Carolina Department of Natural Resources

FI - Fishery Independent FD- Fishery Dependent

52

Table 32.1.2. Summary of age structure of Atlantic croaker obtained from the available data sets

Program Birthdate Type Year Age 0 Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Age 7 Age 8 Age 9 Age 10 Age 11 Age 12 Total

VIMS/ODU 1-Jan FD 1998 0 15 32 29 46 42 23 18 15 1 221

1999 0 3 34 31 19 64 29 57 38 40 2 317

2000 0 1 22 63 36 15 41 25 40 29 37 2 311

2001 0 1 2 19 97 72 17 35 31 45 26 18 1 364

2002 0 10 13 23 63 110 61 12 20 17 17 6 8 360

NC/DMF Not Adjusted FI/FD 1989 0 16 80 96

1990 1 31 32

1996 117 186 204 130 133 55 11 836

1997 35 156 76 47 58 43 12 1 428

1998 271 310 77 87 118 83 97 25 2 1 1,071

1999 111 204 55 32 42 75 76 57 18 1 671

2000 111 249 176 51 46 58 68 27 24 5 815

2001 86 113 151 142 67 43 46 57 50 23 14 1 793

2002 52 154 110 94 103 38 9 12 13 10 5 5 605

SEAMAP 1-Jan FI 2001 361 315 76 41 1 3 0 0 0 0 0 797

2002 338 138 37 21 13 0 1 0 0 0 0 548

VIMS 1-Jul FI/FD ? 1998 0 90 297 238 309 362 420 250 125 8 1 0 2,100

1999 0 26 278 203 146 240 438 396 378 143 9 3 2,260

2000 0 5 107 119 22 39 45 81 29 50 1 1 499

Maryland/SCDNot Adjusted FD 1999 37 38 6 21 10 39 15 13 1 180

2000 3 39 24 9 5 13 21 10 13 8 145

2001 3 29 6 38

53

Table 42.2.1. Summary of Von Bertalanffy Growth parameters examined for use in this assessment. Only studies base on age determination made using sectioned otoliths were considered.

Program NC-DMF VMRC VMRC VMRC Collection Period 1996-2002 1998-2002 1998-2002 1998-2002 1988-1991 ~1450-1765Citation ODU ODU Barbieri et al. 1994b Hales&RietzMethod 1 2 3 4 4 1L infinity 434 558 505 479 312 422K 0.242 0.093 0.135 0.157 0.360 0.180to -1.957 -4.135 -2.713 -3.260 -3.260 -2.360

1 Simple. Straight observed data from age dataset.2 Adjusted for month age, weighted sample size (1/count age group)3 Adjusted for month age, not sample size weighted4 Based on Bio age in months.

54

Table 53.1.1.1: Commercial Landings of Atlantic Croaker in Pounds by Atlantic Coastal States, 1950-2001

YEAR E FL GA SC NC VA MD DE NJ NY RI MA NH TOTAL 1950 60,400 1,000 29,100 2,095,800 6,673,900 2,517,900 6,100 37,900 11,422,100

1951 121,300 22,000 2,102,100 4,223,400 1,850,600 4,900 50,000 8,374,300

1952 151,200 23,000 1,346,300 3,641,200 850,300 8,300 82,700 6,103,000

1953 94,000 6,900 1,433,900 4,060,100 462,400 43,300 156,700 6,257,300

1954 124,700 5,100 1,015,500 5,124,500 912,900 60,100 369,200 7,612,000

1955 201,600 32,200 992,600 9,752,100 1,704,600 667,200 741,300 14,091,600

1956 138,400 73,500 4,828,800 9,667,900 1,748,700 27,200 76,800 16,561,300

1957 131,200 1,700 2,915,900 14,197,600 1,400,000 166,900 103,500 18,916,800

1958 157,600 100 9,700 6,920,600 11,856,000 658,500 3,200 400 19,606,100

1959 85,500 9,000 3,056,600 7,655,400 838,300 8,700 1,800 11,655,300

1960 140,700 300 20,500 2,092,800 3,932,700 586,000 200 8,100 6,781,300

1961 142,700 13,300 1,753,500 3,082,300 48,900 56,900 5,097,600

1962 161,300 600 33,300 1,662,800 1,293,700 11,100 4,300 3,167,100

1963 113,700 700 36,200 2,275,700 122,400 1,500 2,550,200

1964 101,200 400 10,400 1,866,900 394,200 2,400 2,375,500

1965 106,800 2,100 3,400 1,753,400 1,531,700 400 3,397,800

1966 330,700 5,100 1,300 1,267,000 1,463,200 800 3,068,100

1967 143,800 6,000 1,282,800 323,500 1,200 1,757,300

1968 70,000 1,200,800 6,200 100 1,277,100

1969 49,900 1,800 200 1,368,700 63,200 400 1,484,200

1970 66,900 9,400 2,700 806,800 127,900 100 200 1,014,000

1971 89,800 500 1,500 948,200 264,900 200 100 1,305,200

1972 101,100 2,400 400 4,108,600 484,100 500 400 17,700 4,715,200

1973 102,900 14,900 3,100 4,324,100 1,358,300 37,300 37,100 100 5,877,800

1974 65,100 8,500 39,900 6,081,700 1,501,700 120,300 45,100 7,862,300

1975 61,500 4,000 3,500 10,251,700 4,721,300 639,700 1,300 885,100 16,568,100

55

1976 78,400 13,600 1,300 15,038,000 5,897,600 1,069,100 2,600 700,600 100 22,801,300

1977 49,500 7,000 600 18,994,800 8,600,600 692,300 8,900 1,478,600 400 29,832,700

1978 39,470 563 730 19,945,471 8,099,100 597,000 7,300 654,900 100 29,344,634

1979 38,646 19,137 7,082 20,558,193 2,136,600 97,400 3,700 91,000 6,200 2,600 22,960,558

1980 50,911 4,721 5,438 21,146,798 711,600 7,100 12,000 900 21,939,468

1981 72,112 1,038 2,441 11,205,342 429,800 2,100 23,500 200 11,736,533

1982 95,357 2,177 386 10,824,953 119,300 7,000 100 11,049,273

1983 81,737 1,097 3,200 7,249,680 150,400 500 200 200 7,487,014

1984 131,375 3,793 9,170,160 817,700 27,100 57,700 3,000 100 10,210,928

1985 115,641 1,256 8,695,544 2,171,821 9,500 100 48,800 400 11,043,062

1986 177,414 924 9,424,828 2,367,000 137,500 500 106,000 12,214,166

1987 217,932 553 698 7,289,191 2,719,500 119,300 800 357,600 10,705,574

1988 140,242 304 2,614 8,434,415 1,749,200 98,700 200 30,100 10,455,775

1989 96,534 1,950 6,824,088 947,300 89,500 137,100 8,096,472

1990 104,402 32 1,190 5,769,512 198,195 3,584 644 20 6,077,579

1991 56,761 3,436,960 164,126 6,183 700 31,292 10 3,696,032

1992 73,369 210 2,796,612 1,339,388 10,685 800 51,600 4,272,664

1993 51,465 3,267,652 5,264,974 158,062 2,500 183,414 8,928,067

1994 96,018 4,615,791 5,773,430 218,744 3,000 117,256 10,824,239

1995 22,879 6,021,326 6,991,044 549,716 13,000 334,654 13,932,619

1996 26,045 9,961,862 9,442,959 810,435 621,889 1 20,863,191

1997 36,572 10,711,704 12,790,922 1,455,707 10,509 1,994,446 1,309 27,001,169

1998 26,418 10,865,928 12,006,988 1,375,646 10,368 1,029,332 31 25,314,711

1999 26,441 10,185,535 12,849,954 1,584,412 14,729 2,071,046 2 4 26,732,123

2000 34,441 10,122,634 12,889,406 1,501,655 11,121 2,130,465 285 40 26,690,047

2001 14,857 12,017,459 12,929,191 2,233,160 22,736 1,389,837 315 28,607,555

Grand Total 5,068,939 108,232 415,502 334,328,038 227,081,498 27,257,189 1,110,963 16,311,675 12,343 3,274 700 17,700 611,716,053

56

Table 63.1.1.2 Commercial value of landings by state of Atlantic croaker Year DE E FL GA MD MA NH NJ NY NC RI SC VA Total

1950 1,040 2,099 50 351,283 3,250 103,406 1,455 1,210,225 1,672,8081951 783 12,130 264,763 3,343 112,531 1,100 655,990 1,050,6401952 1,238 14,969 155,614 16,540 66,325 920 424,816 680,4221953 5,198 10,340 76,162 20,095 69,118 276 402,822 584,0111954 4,212 13,717 116,446 29,732 50,593 204 508,383 723,2871955 43,456 20,979 200,107 62,545 53,636 3,864 798,522 1,183,1091956 2,197 15,224 238,479 9,770 289,728 7,350 801,002 1,363,7501957 18,430 15,744 134,390 12,304 219,543 89 1,541,111 1,941,6111958 384 15,760 9 72,273 62 530,542 499 1,091,817 1,711,3461959 1,324 8,550 172,667 392 228,331 430 1,215,370 1,627,0641960 50 18,291 27 156,437 1,519 158,029 1,005 642,507 977,8651961 18,551 13,980 14,533 143,774 532 564,620 755,9901962 21,455 48 3,014 1,274 145,544 1,332 293,777 466,4441963 17,394 84 385 152,442 1,473 30,420 202,1981964 15,335 48 527 139,066 521 62,899 218,3961965 18,394 248 76 107,913 167 154,090 280,8881966 45,767 609 166 62,549 76 193,703 302,8701967 24,940 480 204 65,101 57,337 148,0621968 14,520 16 59,836 1,290 75,6621969 11,445 191 62 62,089 20 9,567 83,3741970 15,525 954 29 30 37,875 219 15,491 70,1231971 19,578 48 36 14 53,605 143 33,463 106,8871972 18,364 253 105 2,119 45 227,052 27 67,868 315,8331973 23,815 1,570 5,765 7,388 8 372,198 426 160,774 571,9441974 14,150 917 18,477 6,463 600,375 4,027 205,209 849,6181975 317 16,997 559 52,973 64,382 904,219 404 512,906 1,552,7571976 832 25,074 2,149 117,317 21 59,152 1,577,235 238 789,279 2,571,2971977 1,841 16,009 1,606 68,468 123,431 2,076,370 74 110 910,279 3,198,1881978 1,934 13,329 159 147,107 128,001 2,735,282 38 146 1,410,445 4,436,4411979 1,558 11,223 5,562 40,614 27,745 3,236 4,345,433 949 1,424 493,772 4,931,516

57

1980 17,998 1,423 3,474 4,092 418 5,213,755 1,232 212,490 5,454,8821981 28,731 446 612 5,097 90 3,944,643 762 124,866 4,105,2471982 26,672 967 1,191 17 4,031,186 122 49,441 4,109,5961983 35,065 513 214 16 47 2,842,139 959 45,353 2,924,3061984 51,200 12,004 17,553 3,191 3,027,015 6 1,345 267,690 3,380,0041985 30 53,754 3,818 357 12,619 2,936,732 429 554,191 3,561,9301986 157 68,578 50,422 37,110 3,088,174 355 576,640 3,821,4361987 260 90,786 185 40,552 112,445 2,956,025 283 1,060,709 4,261,2451988 80 81,586 175 42,482 8,031 3,542,549 1,203 899,327 4,575,4331989 48,001 52,379 49,911 3,380,041 1,044 533,036 4,064,4121990 64,540 24 2,667 150 2,959,259 8 511 110,740 3,137,8991991 245 33,571 5,141 8,653 1,518,888 1 90,735 1,657,2341992 198 49,575 211 5,722 12,504 1,010,646 428,793 1,507,6491993 575 39,029 80,800 39,711 990,961 1,846,467 2,997,5431994 844 36,682 129,508 29,575 1,451,218 2,012,748 3,660,5751995 4,494 17,190 288,575 70,648 2,002,495 2,527,690 4,911,0921996 21,471 291,324 122,339 1 3,642,763 3,345,400 7,423,2981997 2,985 26,309 497,880 401,910 564 4,116,610 3,567,206 8,613,4641998 3,980 20,458 453,055 203,363 23 3,450,044 4,161,655 8,292,5781999 4,896 23,714 482,034 413,019 1 3,120,036 2 3,499,416 7,543,1182000 4,423 39,496 569,224 609,845 112 2,987,064 16 5,598,277 9,808,4572001 6,651 13,568 675,770 371,411 173 3,080,386 3,126,152 7,274,111

Total 114,612 1,397,642 19,515 6,096,790 394 2,119 3,122,060 7,817 81,042,369 1,094 36,722 49,898,776 141,739,910

59

Table 73.2: Summary of current regulations for Atlantic croaker

State/Agency Recreational Commercial Other New York None none New Jersey none none Trawling prohibited from 0-2 miles from

shore Delaware 8” none Maryland 9”,25fish

limit 9” Trawling restricted in Ches. Bay; closed

1/1-3/15 PRFC 25per

person/day

Virginia none none Trawling prohibited in state waters North Carolina

Flynets excluded south of C. Hatteras and mesh size restrictions; culling panels required in long haul seines; TEDs required in flounder trawls in most state waters; TED/BRD requirements and minimum mesh restrictions in shrimp trawls.

South Carolina

none none Gear-related restrictions; TED/BRD requirements; license to land/sell

Georgia 8” 25 fish limit

8” 25 fish limit

BRD requirement; no trawling in sounds

Florida none None Net ban in state waters

60

Table 85.1.2.1 Percent landings by gear for Atlantic coast commercial Atlantic croaker harvest

Year Gill Net Haul Seine Trawl Pound Net Total 1950 2.4 59.6 17.6 20.1 99.7

1951 2.5 52.2 29.2 15.8 99.7

1952 2.5 45.4 26.8 24.9 99.7

1953 1.4 36.0 34.6 27.1 99.0

1954 3.2 39.1 27.7 29.5 99.6

1955 6.4 40.7 26.8 26.0 99.8

1956 3.1 29.7 41.6 25.0 99.4

1957 4.1 39.5 26.7 29.3 99.6

1958 2.7 37.3 38.6 20.8 99.4

1959 4.6 44.0 25.3 25.7 99.6

1960 9.8 44.1 27.5 18.0 99.4

1961 4.4 30.8 27.0 35.9 98.1

1962 7.0 33.7 39.4 18.9 99.1

1963 5.2 21.6 65.9 6.6 99.3

1964 5.5 21.5 58.8 13.6 99.4

1965 6.8 19.0 41.6 32.2 99.6

1966 12.7 23.4 40.3 22.1 98.4

1967 9.3 23.0 56.8 9.5 98.6

1968 7.0 15.6 74.6 1.7 99.0

1969 3.4 12.1 82.1 1.7 99.3

1970 7.4 18.7 63.8 8.6 98.4

1971 7.8 24.1 55.4 12.0 99.4

1972 6.1 17.1 70.2 6.0 99.4

1973 11.5 49.8 31.4 7.0 99.7

1974 6.8 47.2 35.6 10.2 99.8

1975 5.0 41.7 38.1 15.1 99.9

1976 7.6 22.3 48.4 21.5 99.8

1977 10.0 29.1 43.0 17.9 99.9

1978 8.5 26.0 45.5 19.8 99.8

1979 10.3 42.0 39.9 7.6 99.8

1980 17.7 37.6 30.7 13.7 99.8

1981 11.7 47.3 21.6 18.8 99.3

1982 11.5 43.4 25.1 19.1 99.1

1983 12.4 57.6 17.6 11.3 98.9

1984 25.6 34.1 25.3 13.7 98.7

1985 25.6 32.4 24.2 16.8 98.9

1986 31.3 36.0 22.4 8.9 98.5

1987 28.4 31.3 20.6 17.6 98.0

1988 31.0 32.0 20.5 15.0 98.5

1989 22.0 43.2 22.8 10.7 98.8

1990 16.1 64.4 8.0 9.8 98.3

1991 26.4 54.1 12.4 5.6 98.4

61

1992 41.2 32.6 20.3 4.1 98.1

1993 42.9 14.6 25.3 16.2 99.0

1994 43.2 11.0 29.9 14.1 98.2

1995 35.3 10.4 31.1 19.2 96.1

1996 39.0 9.5 30.1 17.4 96.1

1997 29.8 12.1 40.0 15.1 96.9

1998 42.3 9.4 26.9 17.5 96.2

1999 29.3 10.7 35.8 23.6 99.5

2000 39.8 8.2 33.4 17.8 99.2

2001 39.3 8.2 30.2 21.8 99.4

2002 15.9 31.3 35.3 16.5 99.0

62

Table 95.1.2.2 Percent Commercial landings of Atlantic croaker, by state, 1950 - 2001 for Atlantic coast states

DE E FL MD NJ NC SC VA

1950 0.1 0.5 22.0 0.3 18.3 0.3 58.4

1951 0.1 1.4 22.1 0.6 25.1 0.3 50.4

1952 0.1 2.5 13.9 1.4 22.1 0.4 59.7

1953 0.7 1.5 7.4 2.5 22.9 0.1 64.9

1954 0.8 1.6 12.0 4.8 13.3 0.1 67.3

1955 4.7 1.4 12.1 5.3 7.0 0.2 69.2

1956 0.2 0.8 10.6 0.5 29.2 0.4 58.4

1957 0.9 0.7 7.4 0.5 15.4 0.0 75.1

1958 0.0 0.8 3.4 0.0 35.3 0.0 60.5

1959 0.1 0.7 7.2 0.0 26.2 0.1 65.7

1960 0.0 2.1 8.6 0.1 30.9 0.3 58.0

1961 0.0 2.8 1.0 1.1 34.4 0.3 60.5

1962 0.0 5.1 0.3 0.1 52.5 1.1 40.8

1963 0.0 4.5 0.1 0.0 89.3 1.4 4.8

1964 0.0 4.3 0.1 0.0 78.6 0.4 16.6

1965 0.0 3.1 0.0 0.0 51.6 0.1 45.1

1966 0.0 10.8 0.0 0.0 41.3 0.0 47.7

1967 0.0 8.2 0.1 0.0 73.0 0.0 18.4

1968 0.0 5.5 0.0 0.0 94.0 0.0 0.5

1969 0.0 3.4 0.0 0.0 92.2 0.0 4.2

1970 0.0 6.6 0.0 0.0 79.5 0.3 12.6

1971 0.0 6.9 0.0 0.0 72.6 0.1 20.3

1972 0.0 2.1 0.0 0.0 87.1 0.0 10.3

1973 0.0 1.8 0.6 0.6 73.6 0.1 23.1

1974 0.0 0.8 1.5 0.6 77.4 0.5 19.1

1975 0.0 0.4 3.9 5.3 61.9 0.0 28.5

1976 0.0 0.3 4.7 3.1 66.0 0.0 25.9

1977 0.0 0.2 2.3 5.0 63.7 0.0 28.8

1978 0.0 0.1 2.0 2.2 68.0 0.0 27.6

1979 0.0 0.2 0.4 0.4 89.5 0.0 9.3

1980 0.0 0.2 0.0 0.1 96.4 0.0 3.2

1981 0.0 0.6 0.0 0.2 95.5 0.0 3.7

1982 0.0 0.9 0.1 0.0 98.0 0.0 1.1

1983 0.0 1.1 0.0 0.0 96.8 0.0 2.0

1984 0.0 1.3 0.3 0.6 89.8 0.0 8.0

1985 0.0 1.0 0.1 0.4 78.7 0.0 19.7

1986 0.0 1.5 1.1 0.9 77.2 0.0 19.4

1987 0.0 2.0 1.1 3.3 68.1 0.0 25.4

1988 0.0 1.3 0.9 0.3 80.7 0.0 16.7

1989 0.0 1.2 1.1 1.7 84.3 0.0 11.7

1990 0.0 1.7 0.1 0.0 94.9 0.0 3.3

1991 0.0 1.5 0.2 0.8 93.0 0.0 4.4

63

1992 0.0 1.7 0.2 1.2 65.5 0.0 31.3

1993 0.0 0.6 1.8 2.1 36.5 0.0 59.1

1994 0.0 0.9 2.0 1.1 42.6 0.0 53.3

1995 0.1 0.2 3.9 2.4 43.2 0.0 50.2

1996 0.0 0.1 3.9 3.0 47.7 0.0 45.3

1997 0.0 0.1 5.4 7.4 39.7 0.0 47.4

1998 0.0 0.1 5.4 4.1 42.9 0.0 47.4

1999 0.1 0.1 5.9 7.7 38.1 0.0 48.1

2000 0.0 0.1 5.6 8.0 37.9 0.0 48.3

2001 0.1 0.1 7.8 4.9 42.0 0.0 45.2

64

Table 105.2.1.2 . Number of Intercept Trips in which Atlantic croaker could have been potentially caught but were not caught (zero trips), the number of intercepts where Atlantic croaker were caught (Positive Trips) and the number of Atlantic croaker measured by Region. Mid Atlantic = North Carolina and states north. South= South Carolina and states south. See section 5.2.4. for methods used to identify a Table potential Atlantic croaker trip.

Mid Atlantic

South Atlantic

Zero trips*Positive Trips

Fish Measured Zero trips*

Positive Trips

Fish Measured

1981 1,141 126 403 188 127 2061982 1,277 115 150 488 315 7801983 2,628 456 709 503 247 4701984 1,190 276 704 687 320 6631985 3,387 1,006 1,951 1,096 509 1,0411986 3,019 1,134 4,297 1,282 285 4311987 1,885 702 2,364 1,540 310 4721988 2,274 707 2,257 969 202 2751989 4,098 1,433 2,497 1,031 269 2961990 3,670 847 1,425 410 329 2951991 5,079 1,319 1,463 691 291 1511992 4,707 1,247 1,800 1,171 516 4571993 4,357 1,341 1,916 1,005 231 1131994 6,106 3,092 5,228 1,175 299 1321995 5,895 1,970 2,747 1,217 226 861996 6,391 1,936 2,806 1,482 204 771997 7,071 2,318 3,161 1,684 248 1081998 6,930 2,704 3,405 1,750 478 2651999 5,390 2,855 3,049 2,485 438 2692000 6,040 2,453 3,109 2,407 430 2762001 7,944 2,709 5,133 2,619 364 2842002 6,491 2,849 6,470 2,551 371 142

65

Table 115.2.1.6. Size categories used to determine recreational discard weights Length class at n th percentile Region State Year 10 15 20 25 50

Mid Atlantic Delaware 1986 200 210 210 220 230 1987 220 220 230 230 240 1988 200 210 210 220 250 1989 210 220 220 230 250 1990 190 200 200 200 220 1991 190 190 200 200 220 1992 200 210 210 210 230 1993 210 210 220 220 240 1994 200 200 210 220 240 1995 240 250 250 250 270 1996 270 270 270 280 300 1997 260 270 280 280 300 1998 230 240 240 250 270 1999 230 240 240 240 270 2000 250 250 250 260 280 2001 280 280 290 290 310 Maryland 1981 210 210 220 230 250 1983 170 170 180 180 210 1984 210 220 230 230 260 1985 200 210 210 210 230 1986 270 290 300 300 320 1987 280 300 300 300 310 1988 260 270 270 270 280 1989 210 220 220 230 250 1990 190 200 200 200 220 1991 190 190 200 210 230 1992 250 250 250 250 250 1993 240 250 250 250 270 1994 230 230 240 250 260 1995 250 250 250 260 280 1996 280 280 280 290 310 1997 260 270 290 300 320 1998 250 260 270 280 310 1999 240 250 250 260 290 2000 270 280 290 300 320 2001 290 300 300 310 320 New Jersey 1987 220 220 230 230 240 1991 180 190 190 200 220 1992 200 210 210 210 230 1993 210 210 220 220 240 1994 200 200 210 220 240 1995 210 220 220 230 250 1996 220 220 220 230 270 1997 240 250 250 260 290 1998 250 250 260 260 300 1999 270 270 280 290 310 2000 280 290 300 300 320

66

2001 290 290 300 300 320 North Carolina 1981 180 180 190 190 210 1982 190 190 210 210 240 1983 170 170 180 180 190 1984 190 190 190 200 220 1985 180 180 190 190 220 1986 210 220 220 230 230 1987 180 190 190 190 220 1988 190 190 200 200 210 1989 190 190 190 200 220 1990 180 180 190 190 210 1991 170 180 180 190 210 1992 190 190 190 200 210 1993 180 180 190 190 210 1994 180 180 190 190 210 1995 190 190 200 200 230 1996 220 220 230 240 270 1997 190 200 200 200 220 1998 200 210 210 210 230 1999 200 210 220 220 240 2000 200 210 210 220 240 2001 210 210 220 220 250 Virginia 1981 200 210 220 220 240 1982 170 180 190 190 200 1983 140 140 150 150 170 1984 200 200 200 210 220 1985 200 200 210 210 220 1986 190 200 210 210 230 1987 210 210 220 220 240 1988 190 200 210 220 240 1989 210 220 230 230 250 1990 180 190 200 200 220 1991 180 190 190 190 210 1992 200 200 210 210 230 1993 210 210 220 220 240 1994 200 210 210 210 240 1995 210 210 220 220 240 1996 210 220 220 230 260 1997 230 240 240 250 280 1998 230 240 250 260 300 1999 220 220 220 230 270 2000 240 240 260 260 300 2001 230 240 250 250 290

South Atlantic Florida 1981 210 220 230 230 260 1982 190 200 200 210 230 1983 200 200 210 210 230 1984 180 190 200 210 230 1985 160 170 170 180 210 1986 200 210 210 220 280 1987 200 200 210 210 230 1988 210 230 230 230 240

67

1989 230 230 240 270 310 1990 180 190 190 200 220 1991 220 220 230 230 240 1992 230 230 240 250 270 1993 230 230 230 230 260 1994 220 230 230 240 260 1995 220 220 240 260 320 1996 180 190 200 210 240 1997 220 230 230 240 250 1998 240 240 260 260 300 1999 180 190 190 200 230 2000 220 230 240 240 270 2001 230 230 240 240 270 Georgia 1981 170 180 180 190 230 1982 180 180 190 190 210 1983 170 180 180 190 220 1984 190 190 200 210 230 1985 170 180 180 180 200 1986 180 180 180 190 200 1987 180 190 190 190 210 1988 180 180 190 190 200 1989 180 180 180 180 190 1990 190 190 190 200 220 1991 210 220 220 220 240 1992 180 180 180 190 210 1993 230 230 230 230 250 1994 200 210 210 220 240 1995 200 210 220 220 270 1996 180 190 200 210 240 1997 220 230 230 240 250 1998 160 180 190 200 230 1999 170 180 180 200 250 2000 210 220 230 240 260 2001 220 230 230 240 270 South Carolina 1981 170 180 180 190 230 1982 180 190 190 200 220 1983 200 200 210 210 230 1984 180 190 190 190 200 1985 130 140 140 150 180 1986 200 200 200 210 230 1987 190 200 200 210 230 1988 190 190 200 210 240 1989 180 180 180 190 200 1990 180 190 190 200 220 1991 210 220 220 220 240 1992 180 190 200 210 220 1993 230 230 230 230 250 1994 200 210 210 220 240 1995 200 210 220 220 270 1996 180 190 200 210 240 1997 220 220 230 230 250 1998 210 220 220 220 230

68

1999 170 180 180 200 250 2000 210 220 230 240 260 2001 220 230 230 240 270

69

Table 125.2.2.1. Recreational Landings (Type A+B1 in numbers) of Atlantic croaker

YEAR DE FL GA MD MA NJ NY NC RI SC VA TOTAL1981 3,003 598,897 35,591 0 0 1,054 0 1,043,240 0 165,743 964,014 2,811,5421982 0 1,682,619 169,749 10,452 0 0 0 596,494 0 193,553 273,039 2,925,9061983 0 1,148,227 75,173 108,354 0 0 0 1,620,909 0 60,811 2,154,134 5,167,6081984 0 2,781,742 202,365 211,035 0 0 0 2,147,870 0 588,114 2,047,720 7,978,8461985 0 1,306,954 144,341 21,276 0 0 0 723,934 0 260,266 2,284,334 4,741,1051986 4,694 5,118,553 69,886 123,578 0 0 0 356,741 0 599,442 6,384,967 12,657,8611987 0 2,580,727 44,782 208,487 0 0 0 904,028 0 166,977 3,234,223 7,139,2241988 1,186 685,778 64,093 1,005,452 0 0 0 2,256,128 0 144,057 4,048,690 8,205,3841989 478 359,418 72,598 22,872 0 0 0 2,131,762 0 217,023 2,203,505 5,007,6561990 281 304,063 585,380 100,674 0 0 0 1,063,452 0 346,632 2,374,679 4,775,1611991 37,499 1,030,115 184,436 288,471 0 16,235 0 434,067 0 100,816 4,298,541 6,390,1801992 9,855 754,596 440,185 117,426 0 0 0 723,822 0 74,051 4,524,040 6,643,9751993 19,353 304,067 89,734 805,560 0 2,552 0 755,998 0 32,701 4,990,098 7,000,0631994 5,718 599,031 102,974 1,633,582 0 1,567 0 1,179,736 0 188,521 6,494,691 10,205,8201995 136,865 438,076 100,825 827,184 0 15,185 0 850,605 0 75,423 5,029,708 7,473,8711996 235,389 116,574 61,956 775,115 0 35,037 0 662,240 0 37,465 4,997,022 6,920,7981997 385,586 235,430 64,050 1,053,233 0 342,088 0 661,115 0 118,428 8,066,926 10,926,8561998 391,234 234,360 64,953 1,126,058 1,477 143,404 0 387,425 0 170,528 6,730,182 9,249,6211999 662,724 403,982 104,439 1,209,572 0 357,260 0 442,185 0 54,761 5,881,670 9,116,5932000 517,885 455,871 128,922 2,674,881 0 1,023,442 0 391,057 0 32,333 5,486,159 10,710,5502001 312,005 426,263 21,503 1,319,929 0 1,177,814 0 635,554 0 19,801 9,335,312 13,248,1812002 261,635 177,751 36,496 1,223,385 0 253,473 0 408,943 0 66,409 9,129,061 11,557,153

71

Table 135.2.2.2. Recreational Landings (Type A+B1 in pounds) of Atlantic croaker

YEAR DE FL GA MD MA NJ NY NC RI SC VA Grand Total

1981 2,317 305,547 9,666 0 0 582 0 426,241 0 67,283 535,298 1,346,9341982 0 754,958 45,160 70,276 0 0 0 264,606 0 67,013 455,250 1,657,2631983 0 510,599 25,411 32,055 0 0 0 395,404 0 14,159 486,005 1,463,6331984 0 1,856,599 80,685 86,462 0 0 0 584,660 0 161,663 634,872 3,404,9411985 0 684,451 40,419 17,168 0 0 0 278,213 0 72,781 843,417 1,936,4491986 2,595 2,783,651 21,503 116,541 0 0 0 126,887 0 173,028 2,034,334 5,258,5391987 0 1,005,055 14,949 191,631 0 0 0 352,347 0 64,697 1,306,817 2,935,4961988 827 316,900 20,313 926,397 0 0 0 935,460 0 54,313 2,390,572 4,644,7821989 284 268,335 21,139 19,189 0 0 0 658,569 0 80,580 1,329,681 2,377,7771990 113 127,525 205,352 37,870 0 0 0 347,183 0 123,797 875,430 1,717,2701991 10,970 460,455 54,117 117,210 0 4,266 0 157,660 0 16,171 1,728,019 2,548,8681992 3,294 407,670 132,595 53,557 0 0 0 233,537 0 28,512 1,768,964 2,628,1291993 9,639 180,517 55,605 476,865 0 844 0 282,907 0 18,008 1,993,912 3,018,2971994 2,892 337,475 34,051 991,169 0 818 0 351,231 0 128,307 3,024,117 4,870,0601995 82,863 301,920 20,860 567,149 0 9,515 0 326,135 0 25,386 2,675,378 4,009,2061996 205,527 50,038 21,797 702,035 0 39,101 0 346,500 0 14,481 2,716,759 4,096,2381997 340,198 113,094 26,272 1,117,998 0 278,758 0 309,457 0 53,863 5,522,196 7,761,8361998 293,559 141,755 30,968 1,150,461 1,790 135,733 0 161,116 0 76,824 5,920,432 7,912,6381999 522,202 231,695 32,374 1,024,400 0 301,958 0 212,989 0 26,356 4,969,279 7,321,2532000 483,964 242,912 62,390 2,672,999 0 1,125,729 0 201,310 0 13,457 4,888,906 9,691,6672001 304,125 320,490 7,844 1,278,701 0 1,132,216 0 355,011 0 10,749 7,674,758 11,083,8942002 250,899 117,880 10,619 1,162,279 0 268,424 0 242,187 0 29,345 7,075,127 9,156,760

72

Table 5.2.2.3. Percent standard error (PSE) estimates for MRFSS landings (Type A+B1 weight) by State and Year

Year DE FL GA MD MA NJ NC SC VA1981 100 23.6 38 0 0 99.9 27.7 40 19.61982 0 24.1 20.3 53.1 0 0 26.6 24 671983 0 18.5 23.7 23.5 0 0 43.3 45.5 26.91984 0 27 21.8 56.6 0 0 15.2 24 17.61985 0 20.2 17.2 31.4 0 0 26.6 23.8 12.91986 89.2 32.2 23.3 40.7 0 0 19.8 22.7 11.51987 0 25.6 16.5 39.5 0 0 10.5 23.4 8.91988 46.8 38.3 22.2 34.8 0 0 15.6 26.1 12.41989 60.1 27.9 24.2 41.6 0 0 10 31.1 7.91990 69.3 26.1 19 31.2 0 0 8.4 41.3 13.21991 22.3 16.3 23.2 27.1 0 41.1 8.3 50.3 11.71992 25.9 12 13 20.1 0 0 8.5 22.2 10.81993 29.9 18.5 29.2 15.3 0 63.5 8.1 32.8 10.51994 25.3 16.7 25.3 12.1 0 66.7 6.9 33.8 7.41995 26.7 41.2 47.9 21.5 0 49.8 10.4 51.6 9.91996 21.4 47.3 36.4 20.7 0 54.1 10.9 34 12.11997 13.4 31.7 50.9 17.4 0 62.7 15.6 21.5 12.21998 11.9 27.7 23.3 13.2 100 26.3 11.2 26 10.41999 12.7 20.6 33.3 12.7 0 18.3 12.1 61 11.12000 16.7 16.7 31.1 11 0 17.5 13 35.8 11.52001 14.7 18.9 32.2 10.8 0 11.9 14.4 51.6 7.42002 16.4 21.6 39.7 8.8 0 19.7 16.9 32.3 6.1

73

Table 145.2.2.4. Percentage of recreational landings by area and mode fished and total landings (numbers)

AREA INLAND OCEAN (<= 3 MI) OCEAN (> 3 MI) Grand TotalMODE PARTY/CHARTER PRIVATE/RENTALSHORE PARTY/CHARTER PRIVATE/RENTALSHORE PARTY/CHARTER PRIVATE/RENTAL

1981 0% 30% 9% 1% 25% 25% 0% 11% 2,811,5421982 0% 32% 38% 1% 5% 23% 0% 1% 2,925,9061983 0% 38% 23% 0% 1% 27% 1% 9% 5,167,6081984 6% 51% 11% 1% 7% 19% 2% 2% 7,978,8461985 4% 53% 18% 0% 12% 9% 1% 3% 4,741,1051986 3% 50% 40% 0% 2% 3% 1% 1% 12,657,8611987 1% 46% 32% 0% 10% 5% 0% 7% 7,139,2241988 2% 60% 8% 0% 9% 9% 2% 11% 8,205,3841989 1% 64% 9% 0% 9% 13% 0% 3% 5,007,6561990 1% 85% 8% 0% 4% 2% 0% 0% 4,775,1611991 1% 78% 16% 0% 3% 2% 0% 1% 6,390,1801992 3% 68% 13% 0% 3% 10% 0% 2% 6,643,9751993 7% 68% 15% 0% 2% 4% 0% 3% 7,000,0631994 6% 70% 11% 0% 5% 6% 0% 2% 10,205,8201995 6% 62% 14% 0% 5% 6% 2% 4% 7,473,8711996 4% 68% 9% 0% 6% 5% 1% 6% 6,920,7981997 5% 72% 4% 3% 4% 3% 7% 2% 10,926,8561998 4% 75% 9% 0% 3% 5% 2% 3% 9,249,6211999 5% 67% 7% 1% 7% 4% 2% 6% 9,116,5932000 8% 69% 10% 1% 2% 3% 2% 6% 10,710,5502001 5% 78% 7% 1% 3% 3% 1% 2% 13,248,1812002 5% 83% 5% 0% 3% 2% 1% 1% 11,557,153

74

Table 155.2.2.5. Estimated total recreational effort and targeted croaker trips by region.

Mid Atlantic = North Carolina and states north. South= South Carolina and states south.

TOTAL TRIPS TARGETED CROAKER TRIPS

YEAR Mid AtlanticSouth Atlantic TOTAL Mid Atlantic

South Atlantic TOTAL

1981 3,691,520 3,990,963 7,682,483 2,329,911 661,654 2,991,5651982 3,396,824 7,173,834 10,570,658 1,983,758 1,507,711 3,491,4681983 7,375,083 5,494,638 12,869,720 5,260,926 1,170,281 6,431,2071984 5,051,370 8,559,747 13,611,118 3,223,967 2,487,146 5,711,1131985 5,242,377 8,742,788 13,985,165 3,185,154 2,328,634 5,513,7871986 5,525,108 7,971,801 13,496,910 3,823,570 2,143,462 5,967,0321987 5,695,083 6,685,151 12,380,233 3,385,338 2,242,932 5,628,2701988 6,091,459 7,538,628 13,630,087 3,492,146 1,562,828 5,054,9741989 6,078,078 6,520,230 12,598,309 3,313,537 1,299,427 4,612,9631990 6,399,454 6,619,192 13,018,646 3,648,991 1,407,813 5,056,8041991 8,715,745 8,881,602 17,597,347 5,352,227 1,967,682 7,319,9091992 8,631,710 10,078,812 18,710,522 5,058,166 1,930,748 6,988,9141993 9,602,187 9,037,958 18,640,146 6,047,580 1,701,791 7,749,3711994 10,313,392 12,071,012 22,384,404 7,133,688 2,212,994 9,346,6821995 12,498,094 9,892,382 22,390,476 7,895,199 1,737,344 9,632,5431996 9,941,977 8,760,298 18,702,275 6,578,210 1,643,307 8,221,5171997 11,824,784 10,063,415 21,888,198 7,960,976 2,108,066 10,069,0421998 13,433,755 9,581,675 23,015,431 9,072,299 2,082,834 11,155,1341999 11,529,623 8,022,066 19,551,690 7,029,247 1,890,377 8,919,6242000 17,246,924 12,093,782 29,340,706 10,565,856 2,622,408 13,188,2642001 18,628,548 12,378,025 31,006,573 11,410,046 2,780,393 14,190,4392002 15,316,012 9,794,446 25,110,458 8,956,822 2,154,108 11,110,929

75

Table 16175.2.2.6. Size distribution of Atlantic croaker weighted by landings (numbers) for the Mid Atlantic region of the fishery (North Carolina and north). Size class in 10 mm intervals is the lower bound.

Len. Class (mm) 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Grand Total

110 0 0 1,479 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1,479

120 0 0 1,479 0 0 0 0 0 0 0 0 1,361 0 2,465 1,076 0 0 0 0 0 0 0 6,381

130 0 0 74,734 0 0 0 0 0 9,155 0 0 1,361 610 10,279 1,281 0 0 0 0 0 0 0 97,420

140 0 0 274,218 11,459 0 3,409 0 0 7,302 0 13,433 0 4,707 7,925 12,221 0 0 1 0 0 0 0 334,673

150 4,400 0 525,197 11,459 0 10,066 4,981 13,902 18,559 115,748 28,457 0 6,365 7,035 4,176 0 0 0 0 0 0 0 750,344

160 8,799 0 220,101 0 57,369 33,385 21,825 23,231 10,312 67,933 57,759 2,299 15,850 18,663 7,176 368 0 0 25,218 0 3,151 4,975 578,413

170 23,918 32,710 248,288 24,009 82,299 82,497 51,660 107,760 38,284 71,011 188,873 11,300 16,997 38,514 31,016 37,901 16,499 9,711 33,789 1,796 4,392 16,753 1,169,979

180 144,959 54,725 378,004 122,718 121,415 245,174 59,033 188,215 118,219 182,006 460,616 120,424 65,675 145,755 82,574 42,826 30,784 2,283 43,732 8,224 31,166 31,054 2,679,580

190 148,724 99,291 519,472 242,445 87,095 236,993 119,073 389,466 270,771 256,758 454,443 291,133 115,377 288,105 101,752 42,667 49,348 5,404 42,866 28,191 18,769 56,619 3,864,763

200 203,201 56,307 511,883 342,424 116,101 371,879 140,045 366,330 320,784 429,128 605,325 423,067 404,690 1,197,789 335,160 121,853 112,579 118,189 84,749 133,819 119,977 73,622 6,588,901

210 217,199 86,163 372,048 736,204 477,804 566,758 254,422 545,338 458,211 442,949 539,308 812,534 636,977 660,149 432,857 198,141 395,846 118,615 206,178 79,788 297,576 161,935 8,696,999

220 194,588 92,929 250,575 717,588 553,936 1,253,295 365,301 641,178 461,248 586,088 722,741 933,574 1,154,676 1,016,816 719,861 976,825 497,917 299,843 939,289 341,173 547,893 785,889 14,053,223

230 193,398 100,884 134,317 626,894 362,234 1,479,820 765,461 572,795 474,322 445,670 857,136 720,301 771,088 1,046,206 691,448 329,243 760,478 436,294 446,971 176,411 546,866 491,317 12,429,557

240 214,637 51,022 9,948 477,345 231,785 860,639 631,748 520,906 470,206 225,394 369,385 671,464 756,064 954,411 552,495 420,967 706,550 475,775 385,575 297,446 444,746 615,700 10,344,208

250 199,696 33,549 100,532 224,115 383,703 572,266 514,151 340,383 449,473 273,898 335,653 520,990 708,241 817,058 667,215 350,201 956,924 409,162 579,380 407,216 789,761 931,705 10,565,272

260 101,348 53,774 93,912 208,301 153,080 399,804 400,022 561,341 389,073 163,837 208,604 280,463 592,058 834,254 650,647 452,719 873,216 417,084 504,848 468,666 595,410 780,013 9,182,474

270 76,411 52,490 54,655 180,009 107,998 228,975 266,959 671,262 277,441 51,193 100,729 214,511 384,125 494,897 594,658 672,492 893,081 453,226 458,268 562,988 632,851 917,513 8,346,733

280 47,765 43,501 3,335 229,514 84,364 166,145 208,892 547,361 195,494 46,643 49,369 161,441 325,344 461,552 448,551 565,748 770,479 619,903 695,326 636,560 931,414 636,609 7,875,309

290 50,333 4,655 49,052 51,113 80,845 67,197 145,434 314,094 117,691 19,480 14,216 91,507 125,757 330,430 319,154 465,247 673,212 738,155 627,631 710,950 865,076 846,646 6,707,876

300 41,320 41,896 0 32,978 54,193 76,478 95,570 311,349 63,467 16,365 27,674 51,681 151,409 393,384 430,522 739,503 663,544 915,303 675,238 819,026 1,032,298 748,821 7,382,020

310 17,599 4,655 0 36,827 11,072 43,288 98,553 338,420 63,259 6,947 14,774 16,534 99,441 185,171 145,558 293,532 528,938 538,697 488,055 717,107 925,463 710,390 5,284,278

320 13,199 4,976 0 0 15,259 40,247 30,023 240,622 35,707 96,297 16,833 9,988 95,993 143,594 182,494 235,786 462,506 463,025 291,164 927,771 1,179,379 911,358 5,396,220

330 35,714 9,631 0 34,377 14,062 34,302 87,150 207,575 73,690 17,846 0 16,112 58,001 85,595 195,718 133,185 363,779 702,539 364,425 800,150 831,554 574,053 4,639,456

340 13,199 13,965 13,333 33,822 30,647 10,875 17,994 140,087 16,197 15,860 0 23,098 11,896 57,500 97,152 193,520 586,949 362,485 357,368 619,768 817,470 724,919 4,158,105

350 10,826 4,655 27,887 63,024 1,384 25,485 22,218 47,979 0 1,959 0 0 39,679 16,366 51,600 91,190 365,270 382,159 364,052 613,620 572,727 274,656 2,976,735

360 17,599 0 13,943 0 47 21,663 18,199 77,128 8,237 3,586 2,728 0 22,066 53,098 25,117 159,089 261,046 243,081 238,543 525,195 451,061 267,826 2,409,253

370 0 9,310 0 0 2,768 15,214 16,893 30,965 11,517 0 0 0 8,035 9,998 35,819 47,804 94,019 193,417 122,032 373,433 405,644 232,134 1,609,002

380 0 4,655 0 0 35 13,782 3,712 95,299 0 0 6,759 0 0 4,930 23,709 45,558 145,055 298,436 144,728 219,603 179,781 165,643 1,351,686

390 32,479 0 0 0 0 5,194 7,425 0 0 2,491 0 0 1,219 9,147 12,436 41,843 70,615 161,150 110,021 140,291 109,551 91,294 795,155

400 0 9,310 5,003 0 24 1,594 0 13,902 0 0 0 0 0 19,055 0 10,033 47,735 176,046 76,567 118,836 131,126 103,511 712,742

410 0 0 0 0 0 0 0 4,568 0 0 0 0 0 5,152 0 20,156 33,225 76,196 140,539 131,040 63,507 32,559 506,942

420 0 0 0 0 24 0 0 0 0 0 0 0 0 0 6,102 14,575 29,784 73,193 24,131 91,084 40,946 24,565 304,404

430 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31,355 28,291 23,791 62,331 32,404 23,088 201,259

440 0 0 0 0 0 3,188 0 0 0 0 0 0 0 0 0 1,832 12,690 26,581 18,341 39,211 59,460 5,106 166,409

450 0 0 0 0 0 366 0 0 0 0 0 0 0 0 0 0 14,927 23,897 28,239 22,272 10,131 9,231 109,063

460 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 21,236 0 0 8,338 20,604 6,270 56,448

470 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3,880 12,359 3,554 15,088 17,088 51,969

480 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7,760 0 7,566 43,731 0 59,057

490 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 25,381 0 0 0 3,705 3,635 32,720

510 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11,114 0 11,114

520 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14,819 0 14,819

530 0 0 0 0 0 0 0 0 0 0 0 0 1,219 0 0 0 0 0 0 0 0 0 1,219

540 0 14,928 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14,928

800 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13,979 0 0 0 0 0 13,979

76

Table 185.2.2.7. Size distribution of Atlantic croaker weighted by landings (numbers) for the South Atlantic region of the fishery (South Carolina and south). Size class in 10 mm intervals is the lower bound.

Length Class (mm) 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Grand Total

90 0 0 0 0 0 0 0 0 0 0 0 0 1,083 0 0 2,805 0 0 0 0 0 0 3,889

100 0 0 259 0 0 0 0 0 0 0 10,837 0 0 0 0 0 0 0 0 0 0 0 11,096

110 0 0 0 0 3,521 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3,521

120 0 0 0 24,731 48,518 0 0 0 0 0 0 0 0 0 10,014 5,610 0 0 0 0 0 0 88,874

130 0 2,221 0 24,731 77,540 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 104,491

140 13,565 0 0 0 71,503 0 0 0 0 0 0 1,157 0 0 0 0 0 0 0 0 0 0 86,225

150 13,258 2,221 1,131 44,755 35,327 0 702 0 0 0 5,667 6,439 0 0 0 0 0 0 17,867 0 0 0 127,367

160 6,841 5,834 5,990 57,765 60,153 1,362 2,284 15,563 0 8,382 0 8,753 2,369 11,637 0 5,610 0 8,406 16,243 0 0 0 217,190

170 11,912 10,413 11,137 102,589 43,761 5,596 70,595 2,698 18,295 32,676 3,778 21,815 0 0 0 5,610 0 764 25,988 0 0 0 367,627

180 24,987 176,337 22,353 92,088 54,165 100,967 200,841 18,797 52,877 92,015 0 36,971 0 7,711 8,198 2,805 0 3,057 25,107 3,835 2,342 725 926,177

190 23,886 258,406 25,948 117,308 72,118 93,124 109,739 20,626 66,471 111,111 1,889 50,512 0 13,214 2,049 16,831 9,490 8,816 12,994 8,729 2,342 725 1,026,328

200 11,181 249,650 151,110 177,542 61,839 476,674 96,578 19,176 40,957 202,320 11,334 71,924 1,083 28,181 8,198 14,026 0 18,066 22,403 8,949 2,633 2,174 1,675,997

210 44,606 195,115 151,527 278,609 234,685 753,214 225,550 65,139 13,314 114,918 88,588 90,566 0 24,291 12,296 5,610 2,773 938 11,521 27,819 6,429 5,104 2,352,613

220 26,885 182,699 130,568 242,155 178,480 316,845 435,296 23,167 37,739 175,411 132,070 94,846 24,664 59,130 86,263 14,026 48,251 39,822 35,515 24,429 18,130 22,179 2,348,572

230 74,306 133,437 151,943 232,563 112,746 458,967 650,934 160,188 50,571 165,142 270,935 96,965 84,178 82,895 4,099 28,051 40,363 47,770 31,776 24,985 46,166 15,083 2,964,062

240 109,870 157,821 105,826 249,360 130,568 181,711 362,206 148,896 14,736 61,952 171,010 138,315 42,031 55,380 44,156 16,831 59,879 69,511 28,932 40,413 57,821 12,655 2,259,881

250 63,918 102,527 76,038 392,813 110,514 334,564 135,533 109,283 14,882 65,050 92,366 75,473 33,449 55,380 2,049 16,831 59,343 18,655 56,367 62,516 29,528 16,573 1,923,654

260 18,882 106,805 60,735 351,019 68,903 159,679 58,839 58,534 17,751 57,682 119,211 128,140 25,662 114,614 30,043 8,415 24,526 29,849 30,817 84,588 44,254 21,069 1,620,018

270 28,293 100,863 77,068 278,015 101,448 98,760 91,323 56,919 3,500 77,099 249,260 54,302 51,814 153,485 8,198 16,831 22,554 28,917 20,006 87,336 41,626 21,322 1,668,936

280 85,732 132,904 51,823 213,316 42,880 910,671 30,615 25,044 6,035 26,582 43,350 43,614 64,156 69,820 30,043 11,221 5,546 37,543 35,460 36,040 36,166 50,079 1,988,641

290 37,764 57,576 83,606 205,386 64,089 997,048 143,262 2,738 48,276 23,679 0 98,457 9,953 23,273 40,058 5,610 37,960 18,475 48,644 35,944 18,417 19,977 2,020,192

300 12,588 44,528 44,929 115,381 11,767 25,935 170,968 18,137 47,192 0 54,187 9,800 6,416 20,470 14,113 8,415 21,383 23,878 36,479 31,280 26,312 13,168 757,327

310 9,410 26,597 60,208 98,391 12,448 137,661 0 57,078 84,483 6,617 0 27,377 0 34,910 2,049 8,415 9,490 16,696 3,968 17,563 16,392 27,874 657,627

320 37,764 17,966 20,098 78,372 776 91,446 0 45,296 18,104 2,206 17,532 61,075 11,153 11,637 190,274 5,610 45,478 6,167 32,994 11,209 22,572 3,046 730,774

330 15,704 30,506 10,822 3,367 12,448 323,798 2,105 5,188 0 4,411 0 44,189 1,083 11,637 0 2,805 14,235 9,251 5,622 18,052 22,523 16,573 554,321

340 34,587 20,704 6,139 54,601 0 90,462 0 15,563 18,104 6,617 0 35,387 7,499 34,910 0 5,610 0 6,167 15,241 7,642 29,548 6,718 395,498

350 6,233 19,172 4,683 0 0 90,462 0 10,375 10,701 2,206 0 33,142 19,247 14,931 0 0 0 9,251 3,653 17,333 20,837 1,344 263,570

360 31,470 1,041 4,683 78,372 0 0 0 10,375 0 0 0 0 2,167 0 30,043 0 11,092 15,418 8,103 1,278 10,038 1,344 205,425

370 18,882 0 2,342 58,779 0 0 0 0 0 0 43,350 5,838 0 0 0 0 0 24,670 1,654 35,537 2,342 6,718 200,110

380 6,233 5,378 2,342 0 0 90,462 1,280 1,287 7,685 0 0 11,677 19,247 23,273 22,078 0 5,546 9,251 13,821 11,588 5,947 6,805 243,898

390 0 0 0 213 18,194 0 3,840 3,860 15,369 0 0 0 6,416 0 60,086 0 0 0 7,276 0 0 0 115,254

400 0 0 9,193 0 20,752 0 0 0 18,104 0 0 0 0 23,273 0 0 0 18,502 2,826 6,153 471 2,687 101,961

410 0 5,205 0 0 0 45,231 0 0 6,035 0 0 22,095 12,831 0 0 0 0 0 0 0 2,206 2,687 96,289

420 0 0 0 0 0 0 0 0 37,857 0 0 0 0 0 0 0 0 0 0 0 2,526 4,031 44,414

430 0 0 4,683 0 62,420 0 0 0 0 0 0 0 0 0 10,014 0 0 0 0 0 0 0 77,118

440 0 0 7,025 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11,903 2,318 0 0 21,246

450 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11,588 0 0 11,588

460 0 0 0 0 0 3,245 0 0 0 0 0 0 0 16,473 0 8,415 0 0 0 0 0 0 28,134

510 31,470 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31,470

77

Table 195.2.3.1. Numbers of Atlantic croaker released alive by recreational fishermen (Type B2).

Mid Atlantic =North Carolina and states north and South=South Carolina and states south.

Mid AtlanticSouth Atlantic TOTAL

1981 1049345 227414 12767591982 719083 407245 11263281983 3341899 568556 39104551984 2444691 1020911 34656021985 3226084 1451908 46779921986 2761265 443468 32047331987 2651742 1955970 46077121988 2090314 332921 24232351989 2277220 149097 24263171990 5363671 596726 59603971991 11606571 813394 124199651992 6025611 456338 64819491993 9818050 239025 100570751994 12447451 572275 130197261995 7235644 339686 75753301996 6868219 251077 71192961997 10712099 280013 109921121998 10144925 578741 107236661999 11399028 1142286 125413142000 15731849 694437 164262862001 11068063 590107 116581702002 11287097 504026 11791123

78

Table20 5.2.3.2. Estimates number and weight of recreational discards. Discard weight (pounds) were estimated using seven methods described in the text.

REGION YEAR NUMBERS lf_50 lf_p10 lf_p15 lf_p20 lf_p25 med origMid Atlantic 1981 731,311 23,484 16,631 18,137 19,117 20,141 29,734 38,839

1982 503,358 18,561 12,014 12,300 13,975 15,197 29,629 38,1851983 2,339,328 42,911 29,115 31,890 36,219 36,636 61,577 90,9221984 1,711,283 59,029 45,199 49,336 49,343 53,030 72,412 84,4151985 2,258,259 72,182 48,596 54,064 55,306 61,333 94,895 100,0571986 1,932,886 75,941 49,345 54,538 64,658 65,473 89,965 99,9151987 1,856,219 79,498 58,241 60,720 68,568 69,647 98,363 106,1691988 1,463,220 66,214 45,041 47,020 51,999 54,011 87,774 101,5941989 1,594,057 64,218 41,642 46,015 48,455 50,554 83,427 89,2181990 3,754,570 115,372 67,887 76,622 89,138 91,465 148,276 165,8521991 8,124,600 241,410 176,523 181,190 196,692 203,285 307,729 337,5561992 4,217,927 168,409 138,307 140,633 144,621 152,711 197,063 216,5761993 6,872,636 309,696 237,540 263,832 267,099 271,928 375,221 416,6171994 8,713,218 343,744 255,699 258,630 274,382 288,441 440,129 498,1331995 5,064,950 227,013 155,741 165,720 187,587 190,439 290,582 337,3711996 4,807,752 253,099 189,738 190,875 196,557 203,770 328,735 378,8031997 7,498,469 467,309 308,109 337,249 359,971 384,678 608,293 742,9071998 7,101,449 512,674 319,410 339,843 372,522 406,377 697,554 798,8931999 7,979,319 478,114 335,113 347,338 364,757 387,360 663,574 793,2992000 11,012,294 932,122 611,176 653,687 727,430 739,493 1,227,414 1,322,0532001 7,747,643 573,284 391,245 412,904 452,144 461,452 757,788 869,307

South Atlantic 1981 159,189 5,571 3,324 3,988 4,012 4,562 8,538 13,209

1982 285,072 9,609 6,343 6,816 7,576 7,771 12,977 15,4331983 397,989 14,941 10,553 10,896 12,068 12,169 19,287 24,0151984 714,637 22,504 14,073 17,191 17,760 18,507 27,769 31,3931985 1,016,337 18,718 13,065 13,947 14,007 15,148 29,775 43,4001986 310,429 14,995 8,962 9,273 9,289 9,714 22,299 21,5971987 1,369,179 50,353 29,298 32,840 37,123 37,495 62,278 67,3561988 233,045 9,791 6,433 7,682 7,874 8,439 12,597 15,3981989 104,367 5,623 3,328 3,328 3,396 3,938 7,375 9,0801990 417,708 12,298 8,833 9,648 9,648 10,348 15,853 17,2231991 569,376 24,774 19,391 20,124 22,654 22,654 29,683 35,6591992 319,437 14,664 11,012 11,141 11,989 12,862 19,221 22,3011993 167,318 8,719 7,226 7,226 7,454 7,454 11,174 14,3861994 400,593 20,161 13,173 14,533 14,735 16,003 26,417 33,8481995 237,781 14,861 6,887 7,451 8,619 9,094 23,497 25,0041996 175,754 6,048 2,277 3,223 3,755 3,977 9,162 12,5121997 196,009 9,685 7,332 7,693 7,948 8,478 12,195 15,0701998 405,118 18,872 13,592 14,112 14,719 15,137 25,356 29,0991999 799,600 26,046 16,372 18,092 18,561 19,255 37,462 50,4182000 486,105 27,450 17,733 18,966 20,266 21,615 36,346 44,6302001 413,075 22,714 15,883 17,342 17,861 18,847 30,914 37,489

79

Table 215.2.4.1. Species used to identify a potential Atlantic croaker intercept by state. The species most likely to be associated with an Atlantic croaker target trip were determined using a jaccard type index.

See text for details SPECIES/GROUP NJ DE MD VA NC SC GA FLBLACK SEA BASS X X X XCLEARNOSE SKATE X DUSKY SMOOTH-HOUND X X OYSTER TOADFISH X XPIGFISH X X PINFISH X X RED DRUM X X XROCKFISH X SOUTHERN FLOUNDER X X SOUTHERN KINGFISH X X SPOT X X X X X X SPOTTED SEATROUT X SUMMER FLOUNDER X X X WEAKFISH X X X X X XWHITE PERCH X ATLANTIC CROAKER X X X X X X X XSTINGRAY SPP XSUMMER FLOUNDER SPP X

81

Table22 5.2.4.2. Summary statistics for the negative binomial generalized linear model and log transformed general linear models used to estimate recreational catch rates

Model Type Negative Binomial GLM Log transformed GLMRegion North South North SouthScaled Deviance 211,741 64,706 NA NADegrees of Freedom 130,000 35,000 NA NAMean Square Error NA NA 0.8869 0.4049R-Square NA NA 0.1756 0.0948Response Variable (numbers) catch catch log (catch +1) log (catch +1)Explanatory Variables:

Year <.0001 <.0001 <.0001 <.0001Wave <.0001 <.0001 <.0001 <.0001Mode <.0001 <.0001 <.0001 <.0001Area 0.0002 <.0001 <.0001 <.0001Hours Fished <.0001 <.0001 <.0001 0.2595Contributors <.0001 <.0001 <.0001 <.0001State <.0001 <.0001 <.0001 <.0001

82

Table 235.2.4.3. Estimates of recreational catch rates and 95% confidence intervals for Atlantic croaker in the Mid Atlantic (North Carolina and North) and South Atlantic regions (South Carolina and south)

MODEL Negative Binomial GLM Log Transformed GLMREGION Mid Atl. South Atl. Mid Atl. South Atl.YEAR Lower 95% Estimate Upper 95% Lower 95% Estimate Upper 95% Lower 95% Estimate Upper 95% Lower 95% Estimate Upper 95%

1981 0.1541 0.2349 0.3581 0.4096 0.5860 0.8383 -0.1527 -0.0374 0.0936 0.3398 0.4386 0.54471982 0.1499 0.2282 0.3474 0.4555 0.5740 0.7232 -0.1343 -0.0175 0.1151 0.4087 0.4745 0.54331983 0.4552 0.6738 0.9974 0.2990 0.3814 0.4866 -0.0523 0.0699 0.2079 0.2559 0.3157 0.37821984 0.4302 0.6475 0.9746 0.3795 0.4684 0.5782 0.0367 0.1760 0.3341 0.2765 0.3288 0.38331985 0.2696 0.3968 0.5841 0.4846 0.5766 0.6861 -0.0330 0.0902 0.2291 0.3349 0.3795 0.42551986 0.4185 0.6161 0.9069 0.3699 0.4421 0.5284 0.0352 0.1673 0.3162 0.1449 0.1836 0.22371987 0.4656 0.6904 1.0237 0.2423 0.2876 0.3414 0.1347 0.2821 0.4485 0.0819 0.1165 0.15221988 0.5458 0.8073 1.1941 0.2100 0.2583 0.3177 0.2349 0.3942 0.5741 0.0852 0.1275 0.17151989 0.5863 0.8602 1.2621 0.2649 0.3222 0.3919 0.1389 0.2824 0.4440 0.1249 0.1667 0.21011990 0.4247 0.6254 0.9208 1.1335 1.4417 1.8336 0.1362 0.2807 0.4436 0.7221 0.8060 0.89391991 0.6133 0.8988 1.3172 0.5349 0.6632 0.8223 0.1740 0.3219 0.4885 0.3315 0.3879 0.44671992 0.5421 0.7950 1.1660 0.5961 0.7086 0.8423 0.1676 0.3151 0.4812 0.3882 0.4338 0.48091993 0.6520 0.9567 1.4037 0.2057 0.2511 0.3066 0.2261 0.3813 0.5560 0.0464 0.0862 0.12761994 0.8803 1.2871 1.8818 0.2500 0.3000 0.3599 0.4193 0.5974 0.7978 0.0930 0.1314 0.17101995 0.5839 0.8545 1.2506 0.2374 0.2850 0.3422 0.2237 0.3776 0.5508 0.0224 0.0585 0.09601996 0.5842 0.8547 1.2504 0.1245 0.1497 0.1799 0.1556 0.3009 0.4643 -0.0133 0.0192 0.05281997 0.8431 1.2324 1.8016 0.1930 0.2289 0.2716 0.2520 0.4090 0.5857 0.0370 0.0694 0.10281998 0.9747 1.4242 2.0810 0.2955 0.3458 0.4047 0.3179 0.4831 0.6690 0.1299 0.1630 0.19711999 1.4417 2.1079 3.0819 0.2997 0.3451 0.3974 0.5829 0.7817 1.0055 0.0532 0.0807 0.10882000 1.0377 1.5173 2.2183 0.2370 0.2734 0.3154 0.3829 0.5565 0.7519 0.0486 0.0765 0.10512001 0.9385 1.3708 2.0022 0.1720 0.1989 0.2301 0.2523 0.4091 0.5854 -0.0119 0.0140 0.04062002 0.7780 1.1352 1.6565 0.2258 0.2603 0.3000 0.3505 0.5196 0.7099 0.0126 0.0391 0.0663

83

Table24 5.3.2.4.1 Estimates of catch per tow in numbers and weight for the NMFS trawl survey using the delta-log normal GLM.

Numbers WeightYear CPUE StdErr CPUE StdErr1982 3.894219 2.453887 1.004295 0.68435511983 58.59242 35.75971 9.532017 5.8877581984 307.2359 137.7837 47.19537 20.832761985 140.9434 52.47678 23.84431 9.0971771986 70.75209 30.68178 12.5425 5.1798981987 20.59788 13.66391 4.516468 3.1349291988 14.48075 10.52384 3.482267 1.9907451989 47.52855 29.2228 8.06636 4.2741171990 38.40878 21.77061 5.39454 3.1299351991 51.34846 29.45326 7.770463 3.9889541992 100.6933 62.01083 12.21884 6.4853781993 29.2586 19.75414 4.332819 3.0738441994 228.9248 136.2203 29.3999 15.690741995 299.7379 133.462 41.69413 18.701531996 210.4528 98.66762 37.87971 17.756331997 70.35722 41.53354 13.38749 7.8149341998 444.6383 204.7653 78.14807 35.387221999 1164.209 467.5473 182.4884 76.789622000 260.3665 112.7817 52.49517 23.692172001 282.5288 132.7705 61.3016 27.847472002 875.687 354.9882 162.6298 62.05909

84

Table 255.3.3.5.1 Spring Atlantic Croaker (Recruit) Indices. Estimates of catch per tow in numbers for the VIMS trawls survey (spring). Data courtesy of VIMS.

Year Geo. 95% C.I.'s C.V. Geo. 95% C.I.'s C.V. N Bay & River N River Only N

Mean Mean (BRI)

1955 0.31 0.17-0.45 20.15 0.45 0.3-0.61 14.47 20

1956 3.28 1.2-7.3 22.81 4.92 2.05-10.48 18.66 48

1957 13.62 0.11-191.83 48.08 11.70 0.15-139.59 47.30 28

1958 0.30 0-0.88 71.25 0.40 0-1.22 68.83 59

1959 0.04 0-0.88 46.61 0.04 0.01-0.07 41.19 48

1960 0.24 0-0.6 57.76 0.35 0-0.97 62.28 54

1961 0.36 0-1.05 67.92 0.24 0-0.62 63.83 28

1962 0.79 0.56-1.05 11.74 0.67 0.47-0.91 12.66 28

1963 0.01 0-0.04 86.67 0.01 0-0.03 70.15 28

1964 0.35 0.16-0.57 25.21 0.32 0.18-0.48 20.50 55

1965 4.01 1.98-7.4 16.06 2.93 1.58-4.98 15.33 48

1966 0.00 0-0.01 -332.05 0.00 0-0.01 100.00 66

1967 0.34 0.19-0.5 19.83 0.26 0.15-0.38 19.42 83

1968 0.11 0.03-0.2 35.79 0.07 0.02-0.14 39.09 87

1969 0.26 0.15-0.39 20.62 0.18 0.1-0.26 21.44 91

1970 0.06 0-0.12 52.38 0.03 0-0.06 49.09 92

1971 0.23 0.12-0.34 21.94 0.15 0.08-0.24 24.38 228

1972 4.37 0-31.89 53.90 3.63 0-24.42 55.62 210

1973 0.12 0.09-0.16 14.60 0.09 0.07-0.13 14.98 417

1974 2.04 1.2-3.19 14.45 1.68 1.03-2.54 14.09 241

1975 2.63 1.64-3.98 12.28 2.00 1.29-2.94 12.40 334

1976 1.08 0.84-1.37 8.65 0.78 0.6-0.97 9.00 591

1977 0.15 0.1-0.2 16.42 0.11 0.06-0.15 20.39 530

1978 0.08 0.05-0.11 16.61 0.05 0.03-0.07 17.94 413

1979 2.18 1.44-3.14 11.43 1.30 0.9-1.79 11.44 119 2.06 117

1980 0.52 0.39-0.66 10.98 0.44 0.34-0.55 10.12 152 1.85 137

1981 0.07 0.04-0.1 19.67 0.07 0.04-0.1 20.36 140 0.24 132

1982 0.11 0.07-0.14 14.68 0.11 0.07-0.14 15.05 168 1.23 148

1983 6.59 4.94-8.71 6.06 6.67 4.98-8.84 6.10 156 9.49 156

1984 1.63 0.83-2.77 18.72 1.61 0.83-2.73 18.59 140 1.23 144

1985 4.98 4.18-5.92 4.05 5.33 4.4-6.42 4.31 106 4.07 106

1986 2.97 2.25-3.84 7.18 3.33 2.52-4.32 7.03 142 3.19 142

1987 4.24 3.47-5.14 4.81 4.24 3.47-5.14 4.80 139 5.47 139

1988 0.32 0.21-0.44 15.52 0.36 0.23-0.49 16.05 234 0.38 234 2.22 84

1989 0.60 0.38-0.85 15.51 0.65 0.41-0.93 15.63 252 0.78 252 4.63 84

1990 0.43 0.23-0.67 21.19 0.48 0.26-0.74 20.56 252 0.52 252 2.98 85

1991 4.41 3.08-6.18 8.36 4.41 3.08-6.18 8.36 307 4.35 238 12.87 83

1992 1.28 0.87-1.78 12.10 1.28 0.87-1.78 12.10 309 1.34 240 10.26 84

1993 2.17 1.5-3.02 10.34 2.17 1.5-3.02 10.34 301 2.21 240 19.40 84

1994 0.90 0.6-1.26 13.54 0.90 0.6-1.26 13.54 300 0.95 240 2.98 84

1995 1.06 0.77-1.39 10.40 1.06 0.77-1.39 10.40 306 0.93 246 5.55 90

1996 0.19 0.11-0.28 19.63 0.19 0.11-0.28 19.63 405 0.16 242 0.36 88

1997 1.47 1.15-1.85 7.78 1.47 1.15-1.85 7.78 419 0.87 255 7.78 100

1998 1.19 0.95-1.47 7.51 1.19 0.95-1.47 7.51 374 0.48 214 6.21 96

1999 1.50 1.05-2.05 10.83 1.50 1.05-2.05 10.83 397 1.28 232 4.08 100

2000 0.60 0.42-0.80 12.68 0.60 0.42-0.80 12.68 413 0.44 245 1.39 97

2001 0.36 0.24-0.49 14.65 0.36 0.24-0.49 14.65 420 0.32 253 1.18 98

2002 1.59 1.07-2.22 11.59 1.59 1.07-2.22 11.59 361 1.10 195 4.59 98

SPRING ATLANTIC CROAKER (RECRUITS) INDICES

Converted Index (RSCI) Unconverted Index (RSI) Original Index

85

Table 265.3.7.1. Mortality estimates for Atlantic croaker based on different studies and methods

Method SourceVirginia MRC/ ODU North Carolina DMF Foster/VIMS Hales&Reitz Barbieri

TypeStratified random commercial samples

Mixed fishery independent/ dependent Stratified random samples ?

Large fish from Commercial

Archeological samples Com & FI

Location Virginia North Carolina Virginia ? Florida MD-NCTime Period 1998-2002 1989-2002 1998-2000 ~1450-1765 1988-1991Max Age 12 12 11 15 8Full -Rec Age 5 1 5 3 2Sample Size 1573 5347 4589 183 1027k 0.2415 0.135431303 0.2415 0.18 0.36Linfinity 505 434 505 422 312.43

Hoenig exp(1.46-1.01Ln(MaxAge) 0.350023244 0.350023244 0.38217593 0.279394448 0.52716802-LN(0.05)/max age 0.249644356 0.249644356 0.272339298 0.199715485 0.374466534-LN(0.01)/max age 0.383764182 0.383764182 0.418651835 0.307011346 0.575646273

Gabriel et al. 3/MaxAge 0.25 0.25 0.272727273 0.2 0.375Alverson and Carney in Deriso and Quninn 3k/(exp(0.38MaxAge*k)-1) 0.36082591 0.475525535 0.415389482 0.301695537 0.543426382Pauly Pauly 0.573010624 0.409413756 0.573010624 0.497050956 0.850728539

86

Table 6.2.1 Deterministic parameter estimates used to develop criteria for sensitivity runs. For each regional model, all possible combinations of these estimates were examined (N=243). Age 1 commercial selectivity was estimated as (1 + estimate of Age 0)/2 in all runs.

.

Steepness Natural SSB initial Commercial Recreational

(h) Mortality (M) Ratio Age 0 Age 0

Selectivity Selectivity

0.6 0.2 0.25 0 0

0.76 0.3 0.75 0.1 0.05

0.85 0.4 1 0.25 0.2

87

Table 27 6.2.1.1. Summary Table of Available fishery Independent and dependent indices

SEAMAP-ALL SEAMAPS SeamapN NMFS

NC CPUE

MRFSS_NO

MRFSS SO VIMS

NCDMF 120

NCDMF195

MD DNR

FL FWCC Trawl

FLFWCC (Seine)

Year weight weight weight weight weight numbers numbers numbers numbers numbers numbers numbers numbers

1973 x x x x x x x 0.12 x x x x x

1974 x x x x x x x 2.04 x x x x x

1975 x x x x x x x 2.63 x x x x x

1976 x x x x x x x 1.08 x x x x x

1977 x x x x x x x 0.15 x x x x x

1978 x x x x x x x 0.08 x x x x x

1979 x x x x x x x 2.18 22.50 x x x x

1980 x x x x x x x 0.52 36.03 x x x x

1981 x x x x x 0.23 0.59 0.07 6.55 x x x x

1982 x x x 1.04 x 0.23 0.57 0.11 20.83 x x x x

1983 x x x 9.92 x 0.67 0.38 6.59 38.95 x 0.40 x x

1984 x x x 48.41 x 0.65 0.47 1.63 21.94 x 0.00 x x

1985 x x x 26.25 x 0.40 0.58 4.98 8.53 x 0.39 x x

1986 x x x 13.57 x 0.62 0.44 2.97 6.30 x 0.83 x x

1987 x x x 4.49 x 0.69 0.29 4.24 8.78 12.12 0.17 x x

1988 x x x 3.35 x 0.81 0.26 0.32 6.40 37.60 x x x

1989 5.19 2.06 16.35 8.02 x 0.86 0.32 0.60 5.18 63.35 0.43 x x

1990 15.91 14.70 15.03 5.39 x 0.63 1.44 0.43 6.50 119.92 0.19 0.43 x

1991 42.51 25.42 79.44 8.05 x 0.90 0.66 4.41 2.99 21.41 0.08 0.27 x

1992 25.61 6.34 150.26 12.95 x 0.80 0.71 1.28 5.67 141.19 0.79 1.03 x

1993 8.72 3.65 26.54 4.26 x 0.96 0.25 2.17 13.60 64.64 1.92 1.00 x

1994 12.95 3.67 65.90 31.01 108.78 1.29 0.30 0.90 8.04 80.82 1.59 0.41 x

1995 11.78 3.00 60.84 44.64 115.90 0.85 0.28 1.06 11.89 52.62 0.74 0.87 x

1996 7.00 1.89 31.91 40.22 176.31 0.85 0.15 0.19 3.47 134.14 1.61 0.93 0.07

1997 4.78 2.54 10.19 13.81 101.94 1.23 0.23 1.47 13.90 85.16 1.65 0.03 0.07

1998 10.93 2.69 59.02 84.25 97.68 1.42 0.35 1.19 28.58 492.92 3.48 x -0.02

1999 10.09 1.52 87.86 192.01 152.70 2.11 0.35 1.50 5.54 133.53 1.43 x 0.05

2000 5.98 1.98 25.63 54.50 126.87 1.52 0.27 0.60 20.89 39.42 1.39 x 0.16

2001 9.93 5.44 21.73 66.94 234.02 1.37 0.20 0.36 5.07 34.91 0.87 x 0.40

2002 6.45 2.13 25.53 170.60 141.27 1.14 0.26 1.59 6.86 x 1.51 x x

88

Table 28 6.2.2.1. Selectivity estimates used in the base age structured production model

Table29 6.2.2.2. Parameter bounds used in the AD model Builder version of the age structured production model .

Selectivity Age 0 Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Age 7 Age 8 Age 9 Age 10

Commercial 0.1 0.55 1 1 1 1 1 1 1 1 1

Recreational 0.05 1 1 1 1 1 1 1 1 1 1

NMFS Survey 1 1 0.59 0 0 0 0 0 0 0 0

MRFSS Survey 0.05 1 1 1 1 1 1 1 1 1 1

SEAMAP Survey 1 0.4 0 0 0 0 0 0 0 0 0

Parameter Upper bound Lower boundVirgin recruitment (R0) 10 25 (log space)Recruitment deviations from S/R curve 7.5 -7.5 (log space)Catchability coefficeints (q) -3 -25 (log space)Fully selected Fishing mortality (by fleet) 0 1.5Steepness (when estimated) 0.2 1

89

Table30 7.1.1. Standardized residuals for the commercial and recreational landings for the Mid Atlantic and South Atlantic base models

(m=0.30, steepness=0.76, SSB initial: virgin ratio=0.75). Mean and standard deviation of the residuals are also included

Com Rec Com Recmean 0.00 0.00 0.00 0.00s.d 0.20 0.06 0.03 0.29N 30 30 30 30

1973 -0.627 -0.769 1.344 0.9921974 -0.169 -0.173 -0.215 -0.0541975 -0.411 -0.383 -0.173 -0.0181976 -0.645 -0.614 -0.175 -0.0191977 -0.950 -0.707 -0.153 -0.0011978 -0.959 -0.651 -0.140 0.0101979 -0.359 -0.192 -0.114 0.0321980 0.949 0.163 -0.126 0.0201981 0.291 0.083 -0.038 0.0621982 0.615 -1.470 0.086 0.2081983 1.759 1.109 0.282 0.2731984 1.202 0.769 1.109 1.4721985 -0.356 -0.062 1.291 0.9161986 -0.880 -0.500 1.678 2.4511987 -1.786 -0.987 1.849 1.1721988 -0.038 -0.276 1.678 0.5861989 0.703 0.761 1.039 0.3731990 -1.130 -0.832 0.747 0.5121991 -0.397 -0.822 -1.004 -1.3931992 -0.708 -1.308 -1.851 -3.1561993 -2.731 -2.280 -1.970 -1.7491994 0.488 0.839 -1.183 -0.6881995 0.444 0.537 -0.603 -0.6831996 0.848 0.684 -1.554 -0.6881997 0.459 0.340 -0.493 -0.2321998 1.302 1.652 0.001 0.2321999 1.073 1.299 -0.146 -0.0302000 0.045 0.124 -0.330 -0.2362001 0.930 1.799 -0.164 0.0972002 1.040 1.864 -0.674 -0.459

South-AtlanticMid-Atlantic

90

Table31 7.1.2. Standardized residuals for the indices used in the base models for the Mid-Atlantic and South Atlantic models

(m=0.30, steepness=0.76, SSB initial: virgin ratio=0.75). Mean and standard deviation of the residuals are also included

NMFS MRFSS SEAMAP MRFSS SEAMAPmean 0.00 0.00 0.00 0.00 0.00s.d. 0.81 0.52 0.65 0.39 0.44N 21 22 14 22 14

1981 -0.3260 -0.19311982 -1.5045 0.0969 -0.17391983 -0.2110 0.9647 -0.58891984 0.9877 -0.2570 -0.22801985 0.4770 -0.5843 -0.17121986 0.1742 -0.0595 -0.13381987 -0.6863 0.3728 -0.05201988 -0.8371 0.8216 -0.47221989 -0.0951 0.7908 0.0408 -0.2367 -0.72241990 -0.5355 0.4215 0.0040 1.2206 0.54931991 -0.5949 0.6114 1.0053 -0.1027 0.91801992 -0.4830 -0.1743 1.5313 0.3146 0.25701993 -1.4764 0.0171 0.1194 -0.2730 -0.13001994 0.1909 0.3560 0.2227 0.2637 0.01541995 0.2287 -0.7161 0.1233 0.4150 0.14151996 0.2498 -0.6671 0.0707 -0.3625 -0.26631997 -0.2011 0.0138 -0.6880 -0.0064 -0.04931998 0.7100 0.2909 -0.5883 0.2752 -0.05141999 1.1204 -0.4350 -0.0048 0.3012 -0.45872000 0.0779 -0.4870 -0.4484 0.3557 -0.26782001 0.7810 -0.4324 -0.7997 0.0241 0.44782002 1.6271 -0.6188 -0.5884 -0.1757 -0.3830

Mid-Atlantic South-Atlantic

91

Table 327.2.1.1. Fully recruited fishing mortality estimate for Atlantic croaker from the base Mid-Atlantic and South Atlantic models.

Rec=recreational fishery; Comm=Commercial fishery. (m=0.30, steepness=0.76, SSB initial: virgin ratio=0.75).

F per Yr Comm Rec Total Comm Rec Total1973 0.06 0.02 0.08 0.14 0.66 0.801974 0.07 0.02 0.09 0.16 0.96 1.121975 0.17 0.05 0.21 0.11 0.63 0.741976 0.29 0.08 0.37 0.15 0.89 1.051977 0.56 0.12 0.67 0.09 0.55 0.651978 0.95 0.18 1.13 0.05 0.30 0.351979 1.26 0.19 1.45 0.06 0.40 0.461980 1.50 0.14 1.64 0.05 0.34 0.391981 1.21 0.08 1.30 0.06 0.26 0.311982 1.50 0.11 1.61 0.07 0.55 0.621983 0.80 0.12 0.92 0.07 0.36 0.431984 0.50 0.05 0.55 0.13 1.12 1.251985 0.46 0.04 0.51 0.12 0.54 0.661986 0.59 0.09 0.68 0.22 1.50 1.721987 0.77 0.09 0.87 0.51 1.50 2.011988 0.78 0.28 1.06 0.33 0.56 0.881989 0.59 0.14 0.73 0.17 0.48 0.651990 0.57 0.09 0.67 0.16 0.54 0.691991 0.25 0.13 0.37 0.07 0.71 0.781992 0.19 0.07 0.26 0.10 1.50 1.601993 0.44 0.09 0.53 0.11 0.74 0.851994 0.27 0.11 0.38 0.28 1.44 1.721995 0.25 0.06 0.31 0.10 1.39 1.491996 0.26 0.05 0.31 0.11 0.33 0.431997 0.39 0.11 0.50 0.10 0.51 0.621998 0.38 0.13 0.51 0.07 0.51 0.581999 0.28 0.06 0.34 0.06 0.63 0.692000 0.24 0.08 0.32 0.10 0.94 1.042001 0.23 0.09 0.31 0.06 1.03 1.092002 0.19 0.07 0.26 0.08 0.39 0.47

Mid-Atlantic South-Atlantic

92

Table33 7.2.1.2. Exploitation rates for Atlantic croaker from the base mid-Atlantic and South-Atlantic models

(m=0.30, steepness=0.76, SSB initial: virgin ratio=0.75)

Mid SouthAtlantic Atlantic

1973 0.06 0.491974 0.08 0.601975 0.17 0.461976 0.27 0.571977 0.43 0.421978 0.60 0.261979 0.69 0.321980 0.72 0.281981 0.65 0.231982 0.72 0.411983 0.53 0.301984 0.37 0.631985 0.35 0.431986 0.43 0.741987 0.51 0.781988 0.58 0.521989 0.46 0.421990 0.43 0.441991 0.27 0.481992 0.20 0.721993 0.36 0.511994 0.27 0.741995 0.23 0.691996 0.23 0.311997 0.35 0.411998 0.35 0.381999 0.25 0.442000 0.24 0.572001 0.23 0.592002 0.20 0.33

93

Table34 7.2.2.1. Population estimates for Atlantic croaker from the base mid-Atlantic and South-Atlantic models.

Age 0 SSB (MT) Age 0 SSB (MT)(millions) MT (millions) MT (millions) MT (millions) MT

1973 85.00 32,134 287.68 68,996 3.90 345 8.12 9061974 93.96 29,287 296.65 69,415 3.34 345 7.55 8801975 93.08 29,214 300.05 69,205 3.35 269 6.74 7241976 93.51 26,600 288.07 63,261 3.02 292 6.63 7551977 92.99 22,068 264.87 52,945 3.14 246 6.23 6631978 91.68 15,516 228.72 38,114 2.90 285 6.38 7331979 87.74 9,481 190.03 24,504 3.08 362 7.02 8941980 59.21 6,498 141.74 16,626 3.24 390 7.33 9591981 36.06 4,446 93.23 11,068 3.46 433 7.88 1,0561982 34.94 3,394 75.57 8,717 3.41 500 8.35 1,1911983 147.18 2,392 178.70 11,930 4.69 438 9.13 1,1251984 55.30 6,963 167.56 17,721 5.22 512 10.79 1,3101985 68.60 8,851 166.23 21,378 5.70 372 10.54 1,0551986 48.10 9,334 144.13 21,502 4.29 469 10.27 1,1891987 46.23 7,859 121.43 18,284 4.21 268 7.94 7671988 55.70 5,861 114.99 14,707 2.59 204 5.79 5631989 49.23 4,722 105.50 12,201 2.36 228 5.28 5891990 47.54 5,057 105.89 12,749 6.30 247 9.17 8081991 117.20 5,322 176.37 16,522 5.84 396 11.41 1,1191992 83.67 9,119 200.10 23,169 2.43 467 8.51 1,0971993 66.30 12,517 197.76 28,891 2.56 211 5.15 5621994 225.02 11,651 334.93 34,402 2.44 212 5.10 5591995 119.34 17,943 339.57 43,439 1.52 139 3.54 3701996 69.02 21,928 285.21 48,148 1.29 99 2.68 2721997 59.11 21,376 231.98 46,155 1.64 137 3.27 3621998 506.56 16,285 630.09 56,808 1.58 155 3.42 3981999 109.53 29,920 526.86 69,382 1.29 168 3.20 4112000 129.52 34,544 453.77 76,195 1.74 151 3.38 3952001 157.16 33,120 429.22 74,926 2.37 130 4.03 3852002 124.07 32,313 390.36 72,032 1.38 146 3.46 376

PopulationMid-Atlantic South-Atlantic

Population

94

Table 357.4.1. Summary of 1000 Monte-Carlo Trials to evaluate uncertainty surrounding the mid-Atlantic model. Estimates from the base mid-Atlantic model are included for comparative purposes.

Percentile 100 97.5 80 75 50 Base Case 25 40 2.5 0Total Likelihood 53.644 51.230 49.705 49.503 48.876 51.509 48.179 48.602 46.737 45.811M 0.40 0.39 0.35 0.34 0.30 0.30 0.26 0.28 0.20 0.20steepness 0.99 0.95 0.84 0.81 0.70 0.76 0.58 0.66 0.34 0.21Comm Age-0 Sel 0.25 0.24 0.19 0.18 0.13 0.10 0.08 0.11 0.02 0.00Rec Age-0 Sel 0.20 0.20 0.15 0.14 0.10 0.05 0.06 0.08 0.01 0.00SSB1/SSB0 1.00 1.00 0.84 0.80 0.65 0.75 0.45 0.58 0.26 0.25R0 21.23 19.51 18.93 18.88 18.69 18.55 18.50 18.62 18.19 17.97NMFS q -12.84 -12.93 -13.03 -13.06 -13.32 -13.22 -13.54 -13.41 -14.45 -16.48MRFSS q -18.05 -18.15 -18.35 -18.40 -18.84 -18.42 -19.18 -19.01 -20.45 -22.33SEAMAP q -11.89 -11.99 -12.12 -12.14 -12.25 -12.21 -12.47 -12.32 -13.25 -15.38Comm F 1973 1.43 0.61 0.09 0.09 0.06 0.06 0.05 0.06 0.03 0.01Comm F 1974 0.38 0.25 0.12 0.11 0.08 0.07 0.06 0.07 0.03 0.01Comm F 1975 0.66 0.42 0.26 0.24 0.18 0.17 0.15 0.16 0.07 0.01Comm F 1976 0.79 0.59 0.41 0.38 0.30 0.29 0.25 0.28 0.11 0.02Comm F 1977 1.16 0.96 0.72 0.69 0.55 0.56 0.46 0.52 0.17 0.03Comm F 1978 1.50 1.50 1.20 1.15 0.93 0.95 0.74 0.85 0.19 0.03Comm F 1979 1.50 1.50 1.50 1.50 1.18 1.26 0.78 1.02 0.15 0.02Comm F 1980 1.50 1.50 1.50 1.50 1.43 1.50 0.78 1.07 0.14 0.02Comm F 1981 1.50 1.50 1.38 1.27 0.85 1.21 0.47 0.66 0.08 0.01Comm F 1982 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 0.09 0.01Comm F 1983 1.50 1.50 1.20 1.13 0.83 0.80 0.59 0.73 0.06 0.01Comm F 1984 0.81 0.72 0.58 0.55 0.38 0.50 0.29 0.34 0.06 0.01Comm F 1985 0.99 0.84 0.65 0.61 0.32 0.46 0.22 0.27 0.05 0.01Comm F 1986 1.42 1.28 1.03 0.94 0.38 0.59 0.23 0.30 0.05 0.01Comm F 1987 1.50 1.50 1.34 1.21 0.39 0.77 0.22 0.30 0.04 0.01Comm F 1988 1.50 1.50 1.38 1.22 0.38 0.78 0.22 0.30 0.04 0.01Comm F 1989 1.49 1.29 0.93 0.83 0.30 0.59 0.17 0.23 0.03 0.01Comm F 1990 1.50 1.50 1.05 0.93 0.28 0.57 0.15 0.21 0.03 0.00Comm F 1991 0.79 0.66 0.46 0.42 0.14 0.25 0.08 0.11 0.02 0.00Comm F 1992 0.46 0.37 0.29 0.27 0.12 0.19 0.08 0.10 0.02 0.00Comm F 1993 1.50 1.50 1.50 1.50 0.27 0.44 0.15 0.21 0.03 0.01Comm F 1994 0.74 0.68 0.58 0.56 0.19 0.27 0.13 0.16 0.03 0.01Comm F 1995 0.49 0.47 0.41 0.40 0.19 0.25 0.14 0.16 0.04 0.01Comm F 1996 0.51 0.47 0.40 0.38 0.22 0.26 0.16 0.19 0.04 0.01Comm F 1997 0.91 0.83 0.70 0.67 0.33 0.39 0.24 0.29 0.06 0.01Comm F 1998 0.92 0.73 0.60 0.57 0.30 0.38 0.21 0.25 0.05 0.01Comm F 1999 0.43 0.40 0.36 0.35 0.22 0.28 0.17 0.20 0.05 0.01Comm F 2000 0.36 0.33 0.30 0.29 0.19 0.24 0.14 0.16 0.04 0.01Comm F 2001 0.32 0.31 0.29 0.28 0.18 0.23 0.13 0.16 0.04 0.01Comm F 2002 0.28 0.27 0.24 0.23 0.15 0.19 0.11 0.13 0.03 0.01

95

Table 7.4.1 continued. Percentile 100 97.5 80 75 50 Base Case 25 40 2.5 0

Rec F 1973 0.17 0.08 0.04 0.03 0.02 0.02 0.02 0.02 0.01 0.00

Rec F 1974 0.14 0.10 0.04 0.04 0.03 0.02 0.02 0.02 0.01 0.00

Rec F 1975 0.21 0.14 0.08 0.07 0.05 0.05 0.04 0.05 0.02 0.00

Rec F 1976 0.27 0.20 0.13 0.12 0.09 0.08 0.07 0.08 0.03 0.01

Rec F 1977 0.35 0.27 0.19 0.17 0.13 0.12 0.10 0.12 0.04 0.01

Rec F 1978 0.49 0.38 0.28 0.27 0.20 0.18 0.15 0.18 0.04 0.01

Rec F 1979 0.47 0.37 0.31 0.30 0.22 0.19 0.15 0.19 0.03 0.00

Rec F 1980 0.29 0.25 0.21 0.21 0.17 0.14 0.09 0.13 0.02 0.00

Rec F 1981 0.15 0.14 0.12 0.12 0.08 0.08 0.05 0.06 0.01 0.00

Rec F 1982 0.22 0.20 0.17 0.16 0.10 0.11 0.06 0.08 0.01 0.00

Rec F 1983 0.28 0.25 0.20 0.19 0.14 0.12 0.10 0.12 0.01 0.00

Rec F 1984 0.14 0.13 0.11 0.10 0.07 0.05 0.05 0.06 0.01 0.00

Rec F 1985 0.08 0.07 0.06 0.06 0.04 0.04 0.03 0.03 0.01 0.00

Rec F 1986 0.19 0.18 0.15 0.14 0.07 0.09 0.05 0.06 0.01 0.00

Rec F 1987 0.25 0.24 0.19 0.17 0.06 0.09 0.04 0.05 0.01 0.00

Rec F 1988 0.89 0.80 0.63 0.57 0.17 0.28 0.10 0.13 0.02 0.00

Rec F 1989 0.41 0.37 0.30 0.27 0.09 0.14 0.05 0.07 0.01 0.00

Rec F 1990 0.27 0.24 0.19 0.17 0.06 0.09 0.03 0.04 0.01 0.00

Rec F 1991 0.43 0.36 0.27 0.25 0.08 0.13 0.05 0.07 0.01 0.00

Rec F 1992 0.20 0.17 0.14 0.13 0.06 0.07 0.04 0.05 0.01 0.00

Rec F 1993 0.26 0.23 0.19 0.18 0.07 0.09 0.04 0.06 0.01 0.00

Rec F 1994 0.37 0.34 0.29 0.28 0.09 0.11 0.06 0.07 0.01 0.00

Rec F 1995 0.16 0.14 0.12 0.12 0.06 0.06 0.04 0.05 0.01 0.00

Rec F 1996 0.11 0.10 0.09 0.08 0.05 0.05 0.03 0.04 0.01 0.00

Rec F 1997 0.27 0.24 0.20 0.20 0.09 0.11 0.07 0.08 0.02 0.00

Rec F 1998 0.37 0.31 0.24 0.23 0.11 0.13 0.07 0.09 0.02 0.00

Rec F 1999 0.14 0.13 0.12 0.11 0.07 0.06 0.05 0.06 0.01 0.00

Rec F 2000 0.14 0.13 0.11 0.11 0.07 0.08 0.05 0.06 0.01 0.00

Rec F 2001 0.16 0.14 0.13 0.12 0.08 0.09 0.06 0.07 0.01 0.00

Rec F 2002 0.14 0.12 0.11 0.10 0.06 0.07 0.05 0.05 0.01 0.00

96

Table 7.4.1 continued.

Percentile 100 97.5 80 75 50 Base Case 25 40 2.5 0

SSB 1973 321,747 70,121 38,671 37,178 29,634 32,134 20,998 26,260 9,607 6,481

SSB 1974 321,747 70,121 38,671 37,178 29,634 29,287 20,998 26,260 9,607 6,481

SSB 1975 316,445 68,828 38,372 36,664 29,817 29,214 22,367 26,781 13,093 9,675

SSB 1976 306,565 65,765 34,897 33,140 27,361 26,600 21,417 24,877 13,938 11,163

SSB 1977 294,606 60,546 29,053 27,588 22,804 22,068 18,419 20,959 13,133 11,117

SSB 1978 281,736 52,202 20,924 19,920 16,275 15,516 13,349 14,974 10,333 9,032

SSB 1979 271,152 44,707 13,442 12,705 9,962 9,481 8,484 9,253 6,954 6,397

SSB 1980 266,476 41,764 10,541 9,825 7,050 6,498 5,963 6,559 5,155 4,666

SSB 1981 257,657 38,320 8,282 7,653 5,426 4,446 4,400 4,710 3,988 3,662

SSB 1982 250,606 35,734 7,447 6,795 4,651 3,394 3,104 3,693 2,539 2,374

SSB 1983 242,039 32,601 3,102 2,947 2,382 2,392 1,861 2,060 1,600 1,466

SSB 1984 289,608 48,631 11,484 10,892 8,396 6,963 5,743 6,614 4,645 4,258

SSB 1985 360,964 59,449 14,791 13,860 10,811 8,851 7,704 8,421 7,051 6,639

SSB 1986 386,312 64,659 16,351 15,291 11,624 9,334 7,706 8,592 6,992 6,638

SSB 1987 377,912 64,859 15,700 14,720 10,441 7,859 5,679 7,021 4,759 4,351

SSB 1988 365,251 63,075 14,494 13,566 8,868 5,861 4,028 5,320 3,221 2,989

SSB 1989 358,172 60,699 13,338 12,271 7,703 4,722 3,282 4,364 2,592 2,394

SSB 1990 351,500 59,799 13,423 12,453 7,943 5,057 3,965 4,852 3,242 2,986

SSB 1991 347,169 60,170 14,073 13,075 8,406 5,322 3,676 4,781 2,777 2,489

SSB 1992 404,299 71,988 18,895 17,978 12,639 9,119 7,134 8,472 5,755 5,265

SSB 1993 463,650 80,615 23,289 22,216 16,257 12,517 10,490 12,038 8,749 7,994

SSB 1994 481,565 84,700 24,047 22,866 15,925 11,651 5,556 7,311 4,679 4,318

SSB 1995 655,621 105,252 32,790 31,408 23,057 17,943 12,859 14,739 11,146 10,232

SSB 1996 744,453 119,305 38,208 36,603 27,290 21,928 16,317 18,465 14,481 13,385

SSB 1997 749,638 123,868 37,601 35,871 26,336 21,376 15,543 18,009 13,333 12,442

SSB 1998 707,321 120,504 31,717 29,896 20,592 16,285 9,942 12,652 7,693 6,695

SSB 1999 1,061,017 158,891 49,968 47,433 35,730 29,920 23,792 26,709 20,688 19,074

SSB 2000 1,184,458 179,638 57,311 54,570 41,606 34,544 28,181 31,556 25,020 23,201

SSB 2001 1,224,356 193,001 58,222 55,585 41,238 33,120 26,595 30,567 23,068 21,933

SSB 2002 1,241,446 200,955 58,839 55,835 41,361 32,313 26,101 30,568 22,389 21,154

97

Table 7.4.1 continued.

Percentile 100 97.5 80 75 50 Base Case 25 40 2.5 0

Age 0 -1973 976 163 107 101 79 85 62 73 40 27

Age 0 -1974 908 158 120 117 101 94 85 95 65 54

Age 0 -1975 907 153 120 117 102 93 85 95 65 54

Age 0 -1976 903 161 122 119 103 94 86 97 67 55

Age 0 -1977 899 161 121 118 102 93 86 96 67 55

Age 0 -1978 917 152 117 114 99 92 84 94 66 55

Age 0 -1979 1,038 139 110 107 96 88 84 91 68 57

Age 0 -1980 438 102 75 72 61 59 50 57 33 21

Age 0 -1981 364 68 38 37 32 36 27 30 21 18

Age 0 -1982 470 62 32 31 27 35 24 26 20 19

Age 0 -1983 3,311 414 255 242 185 147 143 164 116 105

Age 0 -1984 1,862 210 87 76 48 55 35 42 23 17

Age 0 -1985 1,031 123 84 81 71 69 60 66 47 40

Age 0 -1986 741 85 62 59 50 48 41 47 29 25

Age 0 -1987 765 83 57 55 47 46 39 44 29 25

Age 0 -1988 1,007 109 69 67 58 56 49 55 37 32

Age 0 -1989 1,093 114 68 66 56 49 47 53 36 31

Age 0 -1990 1,060 121 60 56 46 48 40 44 33 30

Age 0 -1991 3,291 345 168 157 130 117 115 124 97 87

Age 0 -1992 1,966 210 112 108 92 84 78 87 62 55

Age 0 -1993 1,331 159 83 80 70 66 61 66 50 44

Age 0 -1994 6,156 653 326 309 270 225 235 254 193 171

Age 0 -1995 2,451 267 154 144 117 119 99 110 79 70

Age 0 -1996 1,579 168 93 88 75 69 63 71 50 43

Age 0 -1997 1,317 150 78 74 62 59 54 58 44 38

Age 0 -1998 12,615 1,368 718 675 557 507 483 524 393 364

Age 0 -1999 2,539 269 150 144 118 110 98 110 78 67

Age 0 -2000 3,582 398 189 177 139 130 121 132 97 87

Age 0 -2001 3,997 409 226 211 173 157 147 162 119 105

Age 0 -2002 3,156 318 166 158 132 124 109 122 83 68

98

Table36 8.1.1. Biological Reference Points for Mid-Atlantic region.

Quartiles describe the distribution of Monte-Carlo simulation across varying deterministic inputs. Base model estimates are highlighted in bold. Note that (1-M) SSBmsy estimates for the simulation runs are based on the natural mortality estimate used in the individual

SSBmsy F msy MSY Fmax F40% F35% F30% F2001 SSB01 F avg SSB avg 0.75 Fmsy (1-M) 0.5 SSB

/ F msy /SSB msy / F msy /SSB msy SSB msy msy

Maximum 396,261 0.79 62,974 1.00 0.46 0.58 0.75 1.78 7.12 1.88 6.53 0.59 273,420 198,131

97.5 Percentile 69,520 0.56 14,201 0.84 0.42 0.52 0.66 1.47 4.22 1.62 4.04 0.42 48,083 34,760

75th Percentile 30,303 0.36 10,527 0.65 0.34 0.42 0.53 1.17 2.07 1.24 1.99 0.27 22,140 15,152

Median 23,401 0.28 9,010 0.54 0.30 0.36 0.45 1.04 1.71 1.09 1.65 0.21 16,443 11,701

Base Case 18,783 0.32 8,663 0.54 0.29 0.36 0.44 0.98 1.76 1.02 1.73 0.24 13,148 9,392

25th Percentile 18,676 0.20 8,086 0.44 0.25 0.30 0.37 0.89 1.37 0.94 1.32 0.15 12,745 9,338

2.5th Percentile 12,695 0.08 6,462 0.34 0.20 0.24 0.29 0.30 0.85 0.31 0.81 0.06 8,109 6,348

Minimum 9,992 0.01 1,974 0.31 0.19 0.23 0.28 0.03 0.65 0.04 0.61 0.00 6,095 4,996

Num Trials 1299 1299 1299 1299 1299 1299 1299 1299 1299 1299 1299 1299 1299 1299

99

12.0 Figures

Figure 15.1.2.1 Atlantic coastal commercial landings of Atlantic croaker (metric tons), 1950-2001.

Figure 25.2.2.1. Recreational landings of Atlantic croaker (numbers) by region.

Mid Atlantic includes North Carolina and all states north. South Atlantic includes South Carolina and all states south.

0

2000

4000

6000

8000

10000

12000

14000

16000

‘50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00

year

mt

0

24

6

8

1012

14

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Mid-Atlantic South-Atlantic

100

Figure 35.2.2.2. Recreational landings of Atlantic croaker (pounds) by region.

Mid Atlantic includes North Carolina and all states north. South Atlantic includes South Carolina and all states south.

Figure 45.2.2.3. Recreational landings by area fished and total landings (numbers)

0

2,000,000

4,000,000

6,000,000

8,000,000

10,000,000

12,000,000

14,000,000

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Lan

din

gs (

nu

mb

ers)

INLAND OCEAN (<= 3 MI) OCEAN (> 3 MI)

0

2

4

6

8

10

12

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Mid-Atlantic South-Atlantic

101

Figure 55.2.2.4. Recreational landings by mode fished and total landings (numbers)

Figure 65.2.2.5. Proportion of Atlantic croaker landings by Wave and year.

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001wave 1

wave 2

wave 3

wave 4

wave 5

wave 6

0-0.2 0.2-0.4 0.4-0.6 0.6-0.8

0

2,000,000

4,000,000

6,000,000

8,000,000

10,000,000

12,000,000

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Lan

din

gs (

nu

mb

ers)

PARTY/CHARTER PRIVATE/RENTAL SHORE

102

Figure7 5.2.2.6. Estimated number of total recreational trips and trips targeting Atlantic croaker by region.

Mid-Atl= all states including and north of North Carolina. South-Atl= all states including and south of South Carolina.

Figure 85.2.2.7. Size distribution of Atlantic croaker for the northern region (North Carolina and North).

The circle represent the median length class, the box represent the 25th to 75 percentile and whiskers the 2.5 to 97.5 percentile of size class

0

100

200

300

400

500

600

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Len

gth

Cla

ss (

mm

)

0

5

10

15

20

25

30

35

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Tri

ps (

mill

ions

)

Rec Trips Mid-Atl Rec Trips South-Atl Crk Trips Mid-Atl Crk Trips South-Atl

103

Figure 95.2.2.8. Size distribution of Atlantic croaker for the southern region (South Carolina and South).

The circle represent the median length class, the box represent the 25th to 75 percentile and whiskers the 2.5 to 97.5 percentile of size class

Figure10 5.2.3.1. Ratio of Atlantic croaker released by anglers to those landed.

Mid Atlantic= North Carolina and North. South Atlantic= South Carolina and South.

0

100

200

300

400

500

600

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Len

gth

Cla

ss (

mm

)

0

0.5

1

1.5

2

2.5

1981 1984 1987 1990 1993 1996 1999 2002

Mid-Atlantic South-Atlantic

104

Figure11 5.2.4.1. Recreational catch rates and 95% confidence intervals for Atlantic croaker in the Mid Atlantic region (North Carolina and North) using a negative binomial generalized linear model and log transformed general linear model.

Figure 125.2.4.2. Recreational catch rates and 95% confidence intervals for Atlantic croaker in the South Atlantic region (South Carolina and South) using a negative binomial generalized linear model and log transformed general linear model.

-1

0

1

2

3

4

1981 1984 1987 1990 1993 1996 1999 2002Cat

ch p

er T

rip

( nu

mbe

rs)

Ne gbin GLM Log GLM

0

0.5

1

1.5

2

1981 1984 1987 1990 1993 1996 1999 2002

Cat

ch p

er T

rip

( nu

mbe

rs

Negbin GLM Log GLM

105

Figure13 6.2.1.Normalized estimates for the two major fishery independent indices . NMFS=NEFFC trawl survey, SEAMAP= SEAMAP trawl survey.

Figure 146.2.2 Normalized fishery independent CPUE estimates (by Strata).

Strata 21-67 represent the SEMAP data set from the Atlantic coast of Florida to Cape Hatteras, North Carolina. NMFS strata represent the strata grouped inboxes of one degree latitude. The North Carolina -South Carolina border is between strata 51 and 53. Cape Hatteras, N.C. is between strata 65 and 67.

-1

0

1

2

3

4

1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Nor

mal

ized

CP

UE

(x

-mea

n)/s

td.d

ev

NMFS SEAMAP

21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67N

MF

S35

NM

FS

36N

MF

S37

NM

FS

38N

MF

S39

1989

1991

1993

1995

1997

1999

2001

Strata

Year

0.9-1

0.8-0.9

0.7-0.8

0.6-0.7

0.5-0.6

0.4-0.5

0.3-0.4

0.2-0.3

0.1-0.2

0-0.1

106

Figure 156.2.3. Posterior probability distributions for steepness at varying level of natural mortality used in the core models for the Mid Atlantic region (North). Prior probability distribution based on Myers et al . (2002) is also included.

Figure16 6.2.4. Posterior probability distributions for steepness at varying level of natural mortality used in the preliminary models for the South Atlantic region. Prior probability distribution based on Myers et al . (2002) is also included.

0.000

0.002

0.004

0.006

0.008

0.010

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Steepness

Pro

babi

lity

North_M=.2 North_M=.25 North_M=.3 Prior_prob North_M=.4

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Steepness

Pro

babi

lity

Prior_prob South_M=.2 South_M=.25 South_M=.3 South_M=.4

107

Figure17 6.2.5 Map showing geographical boundaries used to define the mid-Atlantic and south-Atlantic models used in the assessment.

South Atlantic

Mid Atlantic

108

Figure 186.2.1.1. Comparison of Standardized estimates [( obs-mean)/std.dev] for the three major indices used in the Mid-Atlantic model.

-2

-1

0

1

2

3

4

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

SeamapN NMFS

-2

-1

0

1

2

3

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

SeamapN MRFSS_NO

-2

-1

0

1

2

3

4

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

NMFS MRFSS_NO

109

Figure 196.2.1.2. Comparison of Standardized estimates [( obs-mean)/std.dev] of the three major indices used in the mid-Atlantic model to the VIMS spring juvenile index and North Carolina Indices.

-2

-1

0

1

2

3

4

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Stan

dard

ized

Est

imat

es

SeamapN NMFS MRFSS_NO VIMS

-2

-1

0

1

2

3

4

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Sta

nd

ard

ized

Est

imat

es

SeamapN NMFS NCCPUE

MRFSS_NO NCDMF195

110

Figure20 6.2.1.3. Comparison of Standardized estimates [( obs-mean)/std.dev] for the two major indices used in the South-Atlantic model with other available indices for the region.

-2

-1

0

1

2

3

4

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Stan

dard

ized

Est

imat

es

SeamapS MRFSS SO FLFWCCTrawl FLFWCC(Seine)

111

Figure 216.2.2.1 Proportion of commercial landings by size .

Figure22 6.2.2.2. Proportion of commercial landings by estimated age (1973?-2002)

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0 100 200 300 400 500 600

Length Class (mm)

Pro

po

rtio

n o

f L

and

ing

s

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

-3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112131415161718

Estimated Age

Pro

port

ion o

f Landin

gs

112

Figure 236.2.2.3. Predominant size range of Atlantic croaker in the commercial landings overlaid on combined age-length data.

Figure 246.2.2.4. Proportion of recreational landings by size (1981-2002)

0

100

200

300

400

500

600

0 3 6 9 12 15Age

Tot

al L

engt

h (m

m)

0

0.02

0.04

0.06

0.08

0.1

0.12

0 200 400 600 800 1000

Total Length (mm)

Pro

port

ion

of L

andi

ngs

113

Figure25 6.2.2.5 Proportion of recreational landings by estimated age (1981-2002)

Figure 266.2.2.6. Predominant size range of Atlantic croaker in the recreational landings overlaid on combined age-length data.

0.001

0.021

0.041

0.061

0.081

0.101

0.121

-2 -1 0 1 2 3 4 5 6 7 8 9 101112131415161718

Estimated Age

Pro

po

rtio

n o

f L

and

ing

s

0

100

200

300

400

500

600

0 3 6 9 12 15Age

Tot

al L

engt

h (m

m)

114

Figure 276.2.2.7. Proportion of SEAMAP catches by size class

Figure28 6.2.2.8. Proportion of SEAMAP catches by age

0.0

0.1

0.1

0.2

0.2

0.3

0 100 200 300 400

Total Length (mm)

Pro

port

ion

of c

atch

0.0

0.2

0.4

0.6

0.8

1.0

0 1 2 3 4 5 6 7 8

Age

Pro

port

ion

Aged_Len Matrix (01-02) Estimated-Age

115

Figure 296.2.2.9. Predominant size range of Atlantic croaker in the SEAMAP catch overlaid on combined age-length data.

Figure 306.2.2.10. Proportion of NMFS survey catches by size class

0

100

200

300

400

500

600

0 3 6 9 12 15Age

Tot

al L

engt

h (m

m)

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600

Total Length (mm)

Rat

io o

f ca

tch

to

max

Cat

c

116

Figure 316.2.2.11. Proportion of NMFS survey catches by estimated age

Figure 326.2.2.12. Predominant size ranges of Atlantic croaker in the NMFS trawl catch overlaid on combined age-length data.

0

0.2

0.4

0.6

0.8

1

1.2

0 1 2 3 4 5

Est imated Age

0

100

200

300

400

500

600

0 3 6 9 12 15Age

Tot

al L

engt

h (m

m)

117

Figure33 6.2.2.13. Maximum likelihood profile for the prior distribution for the steepness parameter, h, from the covariate analysis of Myers et al.(2002)

0

0.5

1

1.5

2

2.5

0.2 0.4 0.6 0.8 1

Steepness

Pro

babi

lity

dens

ity

118

Figure 347.1.1. Observed and predicted commercial landings from base Mid-Atlantic model

Figure 357.1.2. Observed and predicted recreational landings from base Mid-Atlantic model

Mid-Atlantic

02,0004,0006,0008,000

10,00012,00014,00016,00018,000

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Com

mer

cial

Lan

din

gs (

MT

)

Observed Predicted

Mid-Atlantic

0

1,000

2,000

3,000

4,000

5,000

6,000

1973 1977 1981 1985 1989 1993 1997 2001

Rec

reat

iona

l Lan

ding

s (M

T)

Observed Predicted

119

Figure 367.1.3. Observed and predicted commercial landings from base South-Atlantic model

Figure37 7.1.4. Observed and predicted recreational landings from base South-Atlantic model

South-Atlantic

0

20

40

60

80

100

120

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Com

mer

cial

Lan

ding

s (M

T

Observed Predicted

South Atlantic

0

200400

600

800

1,0001,200

1,400

1,600

1973 1977 1981 1985 1989 1993 1997 2001Rec

reat

ion

al L

and

ings

(M

T)

Observed Predicted

120

Figure 387.1.5. Observed and predicted estimates for the NMFS trawl survey for the base mid-Atlantic model

Figure 397.1.6. Observed and predicted estimates for the MRFSS index for the base mid-Atlantic model

Mid-Atlantic

0

50

100

150

200

250

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

NM

FS

In

d (

Wt/

Tow

)

Observed P redicted

Mid- Atlantic

00.5

11.5

22.5

33.5

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

MR

FSS

Ind

(nu

m/t

rip)

Observed Predicted

121

Figure40 7.1.7. Observed and predicted estimates for the SEAMAP index for the base mid-Atlantic model

Figure 417.1.8. Observed and predicted estimates for the MRFSS index for the base south-Atlantic model

Mid-Atlantic

020406080

100120140160

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

SEA

MA

PIn

dex

(Wt/

Tow

Observed P redicted

South-Atlantic

0

0.5

1

1.5

2

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

MR

FSS

Ind

(nu

m/t

rip)

Obs erved P redic ted

122

Figure 427.1.9. Observed and predicted estimates for the SEAMAP index for the base south-Atlantic model

Figure 437.2.1.1. Fully recruited fishing mortality estimates for Atlantic croaker from the base mid-Atlantic model.

South-Atlantic

0

5

10

15

20

25

30

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

SE

AM

AP

Ind

ex (

Wt/

Tow

)

Obs erved P redic ted

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Fis

hin

g M

orta

lity

Per

Yr

Comme rc ia l Re c re a tiona l Tota l

123

Figure447.2.1.2. Fully recruited fishing mortality estimates for Atlantic croaker from the base South-Atlantic model.

Figure 457.2.2.1. Spawning Stock Biomass and Age 0 estimates for Atlantic croaker from the base mid-Atlantic model.

0.00

0.50

1.00

1.50

2.00

2.50

1973 1977 1981 1985 1989 1993 1997 2001

Fis

hing

Mor

tali

ty P

er Y

r

Comme rc ia l Re c re a tiona l Tota l

0

5

10

15

20

25

30

35

40

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

SSB

(10

00 M

T)

0

100

200

300

400

500

600

Age

0 (

mill

ions

)

SSB Age 0

124

Figure 467.2.2.2. Spawning Stock Biomass and Age 0 estimates for Atlantic croaker from the base south-Atlantic model.

0

100

200

300

400

500

600

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

SS

B (

MT

)

0

1

2

3

4

5

6

7

Age

0 (

Mill

ions

)

SSB Age 0

125

Figure 47487.4.1. Probability profiles used for the deterministic estimates evaluated in the Mote-Carlo sensitivity analysis. For steepness, see Figure 6.2.2.13.

0

0.02

0.04

0.06

0.08

0.2 0.25 0.3 0.35 0.4

Natural Mortality (M per Yr)

Pro

babi

lity

0

0.01

0.02

0.03

0.04

0.25 0.45 0.65 0.85

SSB (1973) :SSB Virgin Constraint

Pro

babi

lity

0

0.02

0.04

0.06

0.08

0 0.05 0.1 0.15 0.2 0.25 0.3

Pro

babi

lity

Rec Age 0 Sel Com Age 0 Sel

126

Figure 497.4.2. Distribution of commercial (A) and recreational (B) fishing mortality rates per year determined using 1,299 Mote-Carlo trails.

The dark horizontal line represents the median value, the box represents the 25th-75th percentiles and the vertical line extends from the 2.5th-97.5th percentile. Each trial represented a unique set of five deterministic input parameters.

Mid-Atlantic

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Com

m F

per

Yr

Mid-Atlantic

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Rec

F p

er Y

r

(B)

(A)

127

Figure50 7.4.3. Distribution of spawning stock biomass estimates determined using 1,299 Mote-Carlo trails.

The dark horizontal line represents the median value, the box represents the 25th-75th percentiles and the vertical line extends from the 2.5th-97.5th percentile. Each trial represented a unique set of five deterministic input parameters. Note: SSB estimates are on log scale.

Mid-Atlantic

1,000

10,000

100,000

1,000,000

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

SSB

( M

T)

128

Figure51 7.4.4. Distribution of Age 0 estimates determined using 1,299 Mote-Carlo trails.

The dark horizontal line represents the median value, the box represents the 25th-75th percentiles and the vertical line extends from the 2.5th-97.5th percentile. Each trial represented a unique set of five deterministic input parameters. Note: SSB estimates are on log scale.

Figure52 8.1.1. Phase plot of the ratio of F2001/ Fmsy with SSB2001/SSBmsy for the Monte-Carlo simulations.

Mid-Atlantic

10

100

1000

10000

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Age

0 R

ecru

its

(mill

ions

)

0.0

0.5

1.0

1.5

2.0

0.0 2.0 4.0 6.0 8.0SSB 01/ SSB msy

F 0

1 /

F m

sy

129

Figure 538.1.2. Phase plot of the ratio of F avg1999-2001/ Fmsy with SSBavg 1999-2001/SSBmsy for the Monte-Carlo simulations.

0.0

0.5

1.0

1.5

2.0

0.0 2.0 4.0 6.0 8.0SSB avg 99-01/ SSB msy

F a

vg 9

9-01

/ F

msy

130

Figure 548.1.3. Estimated fishing mortality rates from the base mid-Atlantic model relative to proposed benchmarks.

Figure 558.1.4. Estimated spawning stock biomass from the base mid-Atlantic model relative to proposed benchmarks.

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Fu

lly

sele

cted

F p

er Y

r

Total F per Yr F msy 0.75 F msy

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

SSB

( M

T)

SSB (1-M) SSB msy 0.5 SSB msy SSB msy

131

Figure 568.2.1 57Beverton and Holt stock recruitment curve and stock recruit scatter for the base mid-Atlantic model. Vertical line represent SSB(MSY)

0

100

200

300

400

500

600

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000

Spawning Stock Biomass (MT)

Age

0 R

ecru

its

( N

umbe

rs in

Mill

ions

)

S-R Curve Estimated R SSB (MSY) Replacement Line

132

Figure58 8.2.2 Beverton and Holt stock recruitment curve and stock-recruit scatter for the base south-Atlantic model. Vertical line represents SSB(MSY)

0

1

2

3

4

5

6

7

0 500 1,000 1,500 2,000 2,500 3,000

Spawning Stock Biomass (MT)

Age

0 R

ecru

its

( Num

bers

in M

illi

ons)

S-R Curve Estimates SSB (MSY) Replacement Line

133

Figure 598.3.1 Yield per recruit and spawning potential ratio curve for the base mid-Atlantic model (m=0.3). Avg represents average SPR from 1999-2002

Figure 608.3.2 Yield per recruit and spawning potential ratio curve for the base south-Atlantic model (m=0.3). Avg represents average SPR from 1999-2002

Mid-Atlantic

0.00

0.05

0.10

0.15

0.20

0.00 0.25 0.50 0.75 1.00 1.25 1.50

Fishing mortality per year

0.00

0.25

0.50

0.75

1.00

Yield per recruit Spawning potential ratio

SPR 30% avg

South Atlantic

0.00

0.05

0.10

0.15

0.20

0.00 0.25 0.50 0.75 1.00 1.25 1.50

Fishing mortality per year

0.00

0.25

0.50

0.75

1.00

Yield per recruit Spawning potential ratio

SPR 30% avg

134

Appendix A.

Comparison of Estimates using the Excel and AD model Builder age structured production model when similarly configured (base Mid-Atlantic model)

Likelihood Components Weight ADMB EXCEL1 1.169 1.5701 0.105 0.134

1.274 1.7052 13.020 12.6971 5.607 5.8182 5.440 5.476

42.527 42.1631 7.734 7.489

51.535 51.357

ADMB EXCEL-13.228 -13.142-18.427 -18.291-12.215 -12.16618.540 18.550

Commercial Landings

Estimated Parameters

MRFSS CPUENEFSC Fall Trawl CPUETotal FleetRecreational Landings

TOTALRecruitment constraintTotal IndicesSEAMAP North CPUE

NEFSC Fall Trawl CPUEMRFSS CPUESEAMAP North CPUEVirgin Recruitment (R0)

ADMB EXCEL ADMB EXCEL ADMB EXCEL1973 0.06 0.06 0.02 0.021974 0.07 0.07 0.02 0.02 -0.122 -0.1291975 0.17 0.17 0.05 0.05 -0.119 -0.1261976 0.29 0.29 0.08 0.08 -0.114 -0.1201977 0.56 0.57 0.12 0.12 -0.103 -0.1091978 0.97 0.98 0.18 0.18 -0.083 -0.0901979 1.28 1.33 0.20 0.20 -0.052 -0.0301980 1.50 1.50 0.14 0.14 -0.306 -0.2341981 1.21 1.15 0.08 0.08 -0.652 -0.7651982 1.50 1.50 0.11 0.11 -0.499 -0.5841983 0.79 0.84 0.12 0.12 1.092 1.0051984 0.49 0.53 0.05 0.06 0.336 0.3091985 0.46 0.53 0.04 0.05 -0.042 -0.0191986 0.58 0.70 0.08 0.10 -0.495 -0.4231987 0.75 0.96 0.09 0.11 -0.554 -0.4751988 0.76 0.96 0.28 0.35 -0.303 -0.2341989 0.58 0.72 0.13 0.16 -0.303 -0.2171990 0.56 0.74 0.09 0.11 -0.229 -0.1791991 0.24 0.31 0.12 0.16 0.641 0.6741992 0.18 0.23 0.07 0.08 0.276 0.3491993 0.43 0.65 0.09 0.11 -0.179 -0.1511994 0.26 0.33 0.11 0.14 0.945 0.9501995 0.25 0.30 0.06 0.07 0.330 0.3521996 0.26 0.29 0.05 0.06 -0.329 -0.3131997 0.39 0.45 0.11 0.13 -0.524 -0.5351998 0.37 0.44 0.13 0.16 1.632 1.6091999 0.28 0.31 0.06 0.06 0.155 0.1512000 0.24 0.27 0.08 0.09 0.213 0.1692001 0.22 0.25 0.09 0.10 0.387 0.3492002 0.19 0.21 0.07 0.08 0.151 0.150

Commercial F per Yr Rec F per yr Rec Deviations

135

Appendix B. Ad model builder template file used in analysesDATA_SECTION

!!USER_CODE ad_comm::change_datafile_name("aspm_dummy.txt"); // define cells init_int nfleets // number of separate fisheries operating init_int firstyr // first year of data considered for each fishery (same for all) init_int lastyr // last year of data considered for each fishery (same for all) init_int firstage // first age in data considered for each fishery (same for all) init_int lastage // last age in data considered for each fishery (same for all) init_int no_ndx // total number of indices // basic fishery inputs init_matrix land(1,nfleets,firstyr,lastyr) init_matrix index(1,no_ndx,firstyr,lastyr) init_matrix sel_fish(1,nfleets,firstage,lastage) init_matrix sel_index(1,no_ndx,firstage,lastage) init_ivector ndx_type(1,no_ndx) // (=1 if weight 0 if numbers) init_vector part_yr(1,no_ndx) // part of year when index takes place // weighting components init_vector ndx_wt(1,no_ndx) init_vector fleet_wt(1,nfleets) init_number ssbc_wt init_number rec_wt init_number prior_wt // biological parameters init_number linf init_number m init_number k init_number to init_number lw_a

136

init_number lw_b init_vector mat_sch(firstage,lastage) // stock recruit inputs init_number ssb_ratio //SSB1/SSB0 constraint init_number h_st //init_number r_init // setting up ranges int ifleet int iyr int iage int indx int i INITIALIZATION_SECTION //steep h_st //init_logr r_init //full_f 0.01 PARAMETER_SECTION number steep //init_bounded_number steep(0.2,1.0,2) //number steep_prior //vector steep_priors(1,5000) //vector steeps(1,5000) init_bounded_vector log_recdev(firstyr+1,lastyr,-7.5,7.5,1) //init_bounded_number init_logr(10,25,1) init_bounded_number log_ro(10,25,1) number init_logr init_bounded_matrix full_f(1,nfleets,firstyr,lastyr,0.0,1.50,1) init_bounded_vector ndx_logq(1,no_ndx,-25,-3,1) matrix len_at_age(firstyr,lastyr,firstage,lastage) matrix wt_at_age(firstyr,lastyr,firstage,lastage) matrix sb_per_r(firstyr,lastyr,firstage,lastage) matrix mat_ogive(firstyr,lastyr,firstage,lastage)

137

vector wgt(firstage,lastage) matrix n_at_age(firstyr,lastyr,firstage,lastage) matrix popwt_at_age(firstyr,lastyr,firstage,lastage) matrix ind_pop_n(1,no_ndx,firstyr,lastyr) matrix ind_pop_wt(1,no_ndx,firstyr,lastyr) matrix pred_totcatch(1,nfleets,firstyr,lastyr) matrix pred_ndx(1,no_ndx,firstyr,lastyr) matrix resid_ndx(1,no_ndx,firstyr,lastyr) matrix resid_catch(1,nfleets,firstyr,lastyr) number ssb_fo 3darray f_by_age(1,nfleets,firstyr,lastyr,firstage,lastage) 3darray c_at_age(1,nfleets,firstyr,lastyr,firstage,lastage) matrix tot_f(firstyr,lastyr,firstage,lastage) matrix z_at_age(firstyr,lastyr,firstage,lastage) vector SSB(firstyr,lastyr) sdreport_vector f_by_yr(firstyr,lastyr); number lik_ndx number lik_catch number lik_ssbc number lik_recc number temp // items used to calculate the std dev of residuals vector obs_ndx(1,no_ndx); vector obs_fleet(1,nfleets); vector sd_res_ndx(1,no_ndx); vector sd_res_fleet(1,nfleets); vector mean_res_indx(1,no_ndx); vector mean_res_fleet(1,nfleets); //MSY stuff swiped from erik's red porgy number avg_land; vector sel_msy(firstage,lastage); matrix N_msy(1,3,firstage,lastage);

138

vector SSB_msy(1,3); //likeprof_number SSB_msy_out; sdreport_number SSB_msy_out; //likeprof_number SdSSB_msy_end; //sdreport_number SdSSB_msy_end; //likeprof_number FdF_msy_end; //sdreport_number FdF_msy_end; //likeprof_number msy_out; vector msy_outx(1,400); vector xx(1,400); sdreport_number msy_out; //likeprof_number F_msy_out; sdreport_number F_msy_out; vector F_msy(1,3); matrix Z_msy(1,3,firstage,lastage); vector L_msy(1,3); vector spr_msy(1,3); vector R_eq(1,3); //sdreport_vector FdF_msy(firstyr,lastyr); //sdreport_vector SdSSB_msy(firstyr,lastyr); number df; number dmsy; number ddmsy; number R0; //likeprof_number profile_steep; likeprof_number profile_ro; objective_function_value f PRELIMINARY_CALCS_SECTION profile_ro.set_stepnumber(20); profile_ro.set_stepsize(0.15); PROCEDURE_SECTION //profile_steep=steep; profile_ro=log_ro; cal_bio_parms(); cal_fmort(); cal_numbers_at_age(); cal_pred_catch(); cal_ndx_abund();

139

cal_pred_index(); cal_msy(); cal_ssqcatch(); cal_ssqndx(); cal_resid_sd(); evaluate_the_objective_function(); FUNCTION cal_bio_parms steep=h_st; //R0=mfexp(log_ro); // create length at age wt at age and ssb_r len_at_age=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { len_at_age(iyr,iage)= linf*(1- mfexp(-k *((iage-1)- to))); } } mat_ogive=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { mat_ogive(iyr,iage)= mat_sch(iage); } } wt_at_age=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { wt_at_age(iyr,iage)= lw_a*pow(len_at_age(iyr,iage),lw_b); } } wgt=0.0; for (iage=firstage;iage<=lastage;iage++) { wgt(iage)=wt_at_age(firstyr,iage); } sb_per_r=0.0;

140

for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=firstage;iage++) { sb_per_r(iyr,iage) = 1; } for (iage=firstage+1;iage<lastage;iage++) { sb_per_r(iyr,iage)= sb_per_r(iyr,iage-1)*mfexp(-m); } for (iage=lastage;iage<=lastage;iage++) { sb_per_r(iyr,iage)= sb_per_r(iyr,iage-1)*(mfexp(-m)/(1-mfexp(-m))); } } ssb_fo=0.0; for (iyr=firstyr;iyr<=firstyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { ssb_fo += sb_per_r(firstyr,iage)*wt_at_age(firstyr,iage)*mat_sch(iage)*0.5; } } FUNCTION cal_fmort for (ifleet=1;ifleet<=nfleets;ifleet++) { f_by_age(ifleet)=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { f_by_age(ifleet,iyr,iage) = full_f(ifleet,iyr)*sel_fish(ifleet,iage); } } } tot_f=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) {

141

for (ifleet=1;ifleet<=nfleets;ifleet++) { tot_f(iyr,iage) += f_by_age(ifleet,iyr,iage); } } } z_at_age=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { z_at_age(iyr,iage) = tot_f(iyr,iage)+m; } } f_by_yr=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=2;iage<=lastage;iage++) { f_by_yr(iyr) += tot_f(iyr,iage); } f_by_yr(iyr) = f_by_yr(iyr)/(nfleets*(lastage-1.0)); } FUNCTION cal_numbers_at_age // note: numbers are normal estimators are in log space // fill the first yr of population n_at_age=0.0; n_at_age(firstyr,firstage)=mfexp(log_ro +log(ssb_ratio)); for (iage=firstage+1;iage<lastage;iage++) { n_at_age(firstyr,iage)=n_at_age(firstyr,iage-1.0)*mfexp(-(m+ tot_f(firstyr,iage-1))); } n_at_age(firstyr,lastage)=n_at_age(firstyr,lastage-1)*(mfexp(-(m+ tot_f(firstyr,lastage-1)))/(1-mfexp(-(m+ tot_f(firstyr,lastage-1))))) ;

142

// ssb for first year SSB = 0.0; for (iage=firstage;iage<=lastage;iage++) { SSB(firstyr) += 0.5 *(n_at_age(firstyr,iage) * mat_ogive(firstyr,iage) *wt_at_age(firstyr,iage)); } //----calculate the age structure each year filling forward and using SSB-R relation for next year's recruits for (iyr=firstyr+1;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=firstage;iage++) { n_at_age(iyr,firstage)=mfexp( log( (0.8*mfexp(log_ro)*steep*SSB(iyr-1)) / (0.2*mfexp(log_ro)*ssb_fo*(1-steep)+SSB(iyr-1)*(steep-0.2))+0.000001 ) + log_recdev(iyr) ); SSB(iyr) += 0.5 * (n_at_age(iyr,iage) * mat_ogive(iyr,iage) * wt_at_age(iyr,iage)); } for (iage=firstage+1;iage<lastage;iage++) { n_at_age(iyr,iage)=n_at_age(iyr-1,iage-1)* mfexp( -(m+ tot_f(iyr-1,iage-1))); SSB(iyr) += 0.5 * (n_at_age(iyr,iage) * mat_ogive(iyr,iage) * wt_at_age(iyr,iage)); } for (iage=lastage;iage<=lastage;iage++) { n_at_age(iyr,lastage)=n_at_age(iyr-1,lastage-1)* mfexp( -(m+ tot_f(iyr-1,lastage-1))) + n_at_age(iyr-1,lastage)* mfexp( -(m+ tot_f(iyr-1,lastage))); SSB(iyr) += 0.5 * (n_at_age(iyr,iage) * mat_ogive(iyr,iage) * wt_at_age(iyr,iage)); } }

143

// population weight at age matrix popwt_at_age=0.0; popwt_at_age = elem_prod(n_at_age,wt_at_age); FUNCTION cal_pred_catch pred_totcatch=0.0; for (ifleet=1;ifleet<=nfleets;ifleet++) { c_at_age(ifleet)=0.0; for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { pred_totcatch(ifleet,iyr)+= (popwt_at_age(iyr,iage)/1000.0) * (f_by_age(ifleet,iyr,iage)/ z_at_age(iyr,iage)) * (1- mfexp(-(z_at_age(iyr,iage)))); c_at_age(ifleet,iyr,iage) = n_at_age(iyr,iage) * (f_by_age(ifleet,iyr,iage)/ z_at_age(iyr,iage)) * (1- mfexp(-(z_at_age(iyr,iage)))); } } } FUNCTION cal_ndx_abund ind_pop_n=0.0; ind_pop_wt=0.0; for (indx=1;indx<=no_ndx;indx++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { ind_pop_n(indx,iyr) += n_at_age(iyr,iage) * sel_index(indx,iage) *mfexp(-1. * part_yr(indx)* (tot_f(iyr,iage)+m)) ; ind_pop_wt(indx,iyr) += popwt_at_age(iyr,iage)*sel_index(indx,iage)*mfexp(-1. * part_yr(indx)* (tot_f(iyr,iage)+m)); } } }

144

FUNCTION cal_pred_index for (indx=1;indx<=no_ndx;indx++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { for (iage=firstage;iage<=lastage;iage++) { if (ndx_type(indx)==1) { pred_ndx(indx,iyr) =ind_pop_n(indx,iyr)*mfexp(ndx_logq(indx)); } else { pred_ndx(indx,iyr) =mfexp(ndx_logq(indx))*ind_pop_wt(indx,iyr) ; } } } } FUNCTION cal_msy //swiped from Erik's red porgy and modified //get ratio of F's from last 3 years not used in final anlyses R0=mfexp(log_ro); df=0.0000001; avg_land=0.0; for (ifleet=1;ifleet<=nfleets;ifleet++) { for (iyr=lastyr-2;iyr<=lastyr;iyr++) { avg_land +=land(ifleet,iyr); } } sel_msy=0.0; for (ifleet=1;ifleet<=nfleets;ifleet++) { for (iage=firstage; iage<=lastage; iage++) { for (iyr=lastyr-2;iyr<=lastyr;iyr++) { sel_msy(iage) +=sel_fish(ifleet,iage)*land(ifleet,iyr)/avg_land;

145

} } } //use Newton's method to get Fmsy, MSY, and Smsy F_msy(1)=0.05; for (i=1; i<=10; i++) { F_msy(2)=F_msy(1)-df; F_msy(3)=F_msy(1)+df; L_msy=0.0; Z_msy(1)=sel_msy*F_msy(1)+m; Z_msy(2)=sel_msy*F_msy(2)+m; Z_msy(3)=sel_msy*F_msy(3)+m; //Initial age N_msy(1,1)=1.0; N_msy(2,1)=1.0; N_msy(3,1)=1.0; for (iage=2; iage<=lastage;iage++) { N_msy(1,iage)=N_msy(1,iage-1)*mfexp(-1.*Z_msy(1,iage-1)); N_msy(2,iage)=N_msy(2,iage-1)*mfexp(-1.*Z_msy(2,iage-1)); N_msy(3,iage)=N_msy(3,iage-1)*mfexp(-1.*Z_msy(3,iage-1)); } //last age is pooled N_msy(1,lastage)=N_msy(1,lastage-1)*mfexp(-1.*Z_msy(1,lastage-1))/(1.-mfexp(-1.*Z_msy(1,lastage))); N_msy(2,lastage)=N_msy(2,lastage-1)*mfexp(-1.*Z_msy(2,lastage-1))/(1.-mfexp(-1.*Z_msy(2,lastage))); N_msy(3,lastage)=N_msy(3,lastage-1)*mfexp(-1.*Z_msy(3,lastage-1))/(1.-mfexp(-1.*Z_msy(3,lastage))); spr_msy(1)=sum(elem_prod(elem_prod(N_msy(1),wgt),mat_sch)); spr_msy(2)=sum(elem_prod(elem_prod(N_msy(2),wgt),mat_sch)); spr_msy(3)=sum(elem_prod(elem_prod(N_msy(3),wgt),mat_sch)); R_eq(1)=(R0/((5*steep-1)*spr_msy(1)))*(4*steep*spr_msy(1)-ssb_fo*(1-steep)); R_eq(2)=(R0/((5*steep-1)*spr_msy(2)))*(4*steep*spr_msy(2)-ssb_fo*(1-steep)); R_eq(3)=(R0/((5*steep-1)*spr_msy(3)))*(4*steep*spr_msy(3)-ssb_fo*(1-steep)); //Initial age N_msy(1)=R_eq(1); N_msy(2)=R_eq(2); N_msy(3)=R_eq(3); for (iage=2; iage<=lastage; iage++) {

146

N_msy(1,iage)=N_msy(1,iage-1)*mfexp(-1.*Z_msy(1,iage-1)); N_msy(2,iage)=N_msy(2,iage-1)*mfexp(-1.*Z_msy(2,iage-1)); N_msy(3,iage)=N_msy(3,iage-1)*mfexp(-1.*Z_msy(3,iage-1)); } //last age is pooled SSB_msy=0.0; N_msy(1,lastage)=N_msy(1,lastage-1)*mfexp(-1.*Z_msy(1,lastage-1))/(1.-mfexp(-1.*Z_msy(1,lastage-1))); N_msy(2,lastage)=N_msy(2,lastage-1)*mfexp(-1.*Z_msy(2,lastage-1))/(1.-mfexp(-1.*Z_msy(2,lastage-1))); N_msy(3,lastage)=N_msy(3,lastage-1)*mfexp(-1.*Z_msy(3,lastage-1))/(1.-mfexp(-1.*Z_msy(3,lastage-1))); SSB_msy(1)=0.5*sum(elem_prod(elem_prod(N_msy(1),wgt),mat_sch)); SSB_msy(2)=0.5*sum(elem_prod(elem_prod(N_msy(2),wgt),mat_sch)); SSB_msy(3)=0.5*sum(elem_prod(elem_prod(N_msy(3),wgt),mat_sch)); L_msy=0.0; for(iage=firstage; iage<=lastage; iage++) { L_msy(1)+=N_msy(1,iage)*((Z_msy(1,iage)-m)/Z_msy(1,iage))*(1.-mfexp(-1.*Z_msy(1,iage)))*wgt(iage); L_msy(2)+=N_msy(2,iage)*((Z_msy(2,iage)-m)/Z_msy(2,iage))*(1.-mfexp(-1.*Z_msy(2,iage)))*wgt(iage); L_msy(3)+=N_msy(3,iage)*((Z_msy(3,iage)-m)/Z_msy(3,iage))*(1.-mfexp(-1.*Z_msy(3,iage)))*wgt(iage); } dmsy=(L_msy(3)-L_msy(2))/(2.*df); ddmsy=(L_msy(3)-2.*L_msy(1)+L_msy(2))/square(df); if(square(ddmsy)<=1e-12) { F_msy(1)=F_msy(1); } if(square(ddmsy)>1e-12) { F_msy(1)-=(dmsy/ddmsy); } if(F_msy(1)<=df) { F_msy(1)=df; } } msy_out=L_msy(1); F_msy_out=F_msy(1); SSB_msy_out=SSB_msy(1);

147

FUNCTION cal_ssqcatch resid_catch=0.0; for (ifleet=1;ifleet<=nfleets;ifleet++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { if(land(ifleet,iyr) > 0.0) { resid_catch(ifleet,iyr)=log(land(ifleet,iyr)) - log(pred_totcatch(ifleet,iyr)); } else { resid_catch(ifleet,iyr)=0.0; } } } FUNCTION cal_ssqndx for(indx=1;indx<=no_ndx;indx++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { if(index(indx,iyr)>0.0) { resid_ndx(indx,iyr)= log(index(indx,iyr))-log(pred_ndx(indx,iyr)); } else { resid_ndx(indx,iyr)=0.0; } } } FUNCTION cal_resid_sd sd_res_ndx=0.0; obs_ndx=0.0; mean_res_indx=0.0; for(indx=1;indx<=no_ndx;indx++) { for (iyr=firstyr;iyr<=lastyr;iyr++)

148

{ if(index(indx,iyr)>0.0) { obs_ndx(indx) += 1.; } } mean_res_indx(indx)= sum(resid_ndx(indx))/obs_ndx(indx); } for(indx=1;indx<=no_ndx;indx++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { sd_res_ndx(indx) += square(resid_ndx(indx,iyr) - mean_res_indx(indx))/ (obs_ndx(indx)-1); } sd_res_ndx(indx) =sqrt(sd_res_ndx(indx) ); } sd_res_fleet=0.0; obs_fleet=0.0; mean_res_fleet=0.0; for (ifleet=1;ifleet<=nfleets;ifleet++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { if(land(ifleet,iyr) > 0.0) { obs_fleet(ifleet) += 1.0; } } mean_res_fleet(ifleet) =sum(resid_catch(ifleet))/obs_fleet(ifleet); } for (ifleet=1;ifleet<=nfleets;ifleet++) { for (iyr=firstyr;iyr<=lastyr;iyr++) {

149

sd_res_fleet(ifleet) += square(resid_catch(ifleet,iyr) - mean_res_fleet(ifleet))/ (obs_fleet(ifleet)-1); } sd_res_fleet(ifleet)=sqrt(sd_res_fleet(ifleet)); } FUNCTION evaluate_the_objective_function lik_ndx=0.0; for(indx=1;indx<=no_ndx;indx++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { lik_ndx +=square(resid_ndx(indx,iyr))*ndx_wt(indx); } } lik_catch=0.0; for (ifleet=1;ifleet<=nfleets;ifleet++) { for (iyr=firstyr;iyr<=lastyr;iyr++) { lik_catch += square(resid_catch(ifleet,iyr))*fleet_wt(ifleet); } } lik_recc=0.0; for (iyr=firstyr+1;iyr<=lastyr;iyr++) { lik_recc += pow(log_recdev(iyr),2)*rec_wt; } //lik_ssbc = ssbc_wt*square((mfexp(init_logr))/(mfexp(log_ro))- ssb_ratio); f += lik_catch + lik_ndx +lik_recc ; RUNTIME_SECTION maximum_function_evaluations 10000 convergence_criteria 1.e-9

150

REPORT_SECTION report <<" ASPM output" << endl; report <<" "<< endl; report <<"input landings"<<endl; report <<" first year "<< endl; report << firstyr <<endl; report <<" "<< endl; report <<" last year "<< endl; report << lastyr <<endl; report <<" "<< endl; report <<" first age "<< endl; report << firstage <<endl; report <<" "<< endl; report <<" last age "<< endl; report << lastage <<endl; report <<" "<< endl; report <<" number of indices "<< endl; report << no_ndx <<endl; report <<" "<< endl; report <<" landings "<< endl; report << land <<endl; report <<" "<< endl; report <<" index values note: -1= no data "<< endl; report << index <<endl; report <<" "<< endl; report <<" fishery selectivities "<< endl; report << sel_fish <<endl; report <<" "<< endl; report <<" index selectivities "<< endl; report << sel_index <<endl; report <<" "<< endl; report <<" index type numbers =0 weight=1 "<< endl; report << ndx_type <<endl; report <<" "<< endl; report <<" index weight for likelihood "<< endl; report << ndx_wt <<endl; report <<" "<< endl; report <<" fleet weight for likelihood "<< endl; report << fleet_wt <<endl; report <<" "<< endl; report <<" ssb ratio weight for likelihood "<< endl; report << ssbc_wt <<endl; report <<" "<< endl; report <<" rec wt for likelihood "<< endl; report << rec_wt <<endl; report <<" "<< endl;

151

report <<" prior wt for likelihood "<< endl; report << prior_wt <<endl; report <<" "<< endl; report <<" l inf "<< endl; report << linf <<endl; report <<" "<< endl; report <<" m "<< endl; report << m <<endl; report <<" "<< endl; report <<" k "<< endl; report << k <<endl; report <<" "<< endl; report <<" to "<< endl; report << to <<endl; report <<" "<< endl; report <<" lw a "<< endl; report << lw_a <<endl; report <<" "<< endl; report <<" lw b "<< endl; report << lw_b <<endl; report <<" "<< endl; report <<" maturity schedule "<< endl; report << mat_sch <<endl; report <<" estimated parameters "<< endl; report <<" "<< endl; report <<" steepness "<< endl; report << steep<< endl; report <<"recruit deviations"<< endl; report << log_recdev << endl; report <<"inital log recruit "<< endl; report << init_logr << endl; report <<" log ro "<< endl; report << log_ro << endl; report <<" full f "<< endl; report << full_f << endl; report <<" log q for indices "<< endl; report << ndx_logq << endl; report <<" calculated parameters "<< endl; report <<" "<< endl; report <<" length at age "<< endl; report << len_at_age << endl; report <<" weight at age "<< endl; report << wt_at_age << endl; report <<" spawning biomass per recruit "<< endl; report << sb_per_r << endl; report <<" maturity ogive "<< endl;

152

report << mat_ogive << endl; report <<" ssb per r at F=0 "<< endl; report << ssb_fo << endl; report <<" population estimates "<< endl; report <<" "<< endl; report <<" numbers at age "<< endl; report << n_at_age << endl; report <<"pop weight at age "<< endl; report << popwt_at_age << endl; report <<" index pop num "<< endl; report << ind_pop_n << endl; report <<" index pop wt "<< endl; report << ind_pop_wt << endl; report <<" predicted total landings "<< endl; report << pred_totcatch << endl; report <<" predicted indices "<< endl; report << pred_ndx << endl; report <<" Residuals "<< endl; report <<" "<< endl;report <<" residuals of index "<< endl; report << resid_ndx << endl; report <<" residuals of catch "<< endl; report << resid_catch << endl; report << " std deviations of residuals" <<endl; report << " " << endl; report << " number of obs for indices residuals " << endl; report << obs_ndx << endl; report << " " << endl; report << " mean of residuals for indices" << endl; report << mean_res_indx << endl; report << " " << endl; report << " std deviation of index residuals " << endl; report << sd_res_ndx << endl; report << " number of obs for fleet residuals " << endl; report << obs_fleet << endl; report << " " << endl; report << " mean of residuals for fleets" << endl; report << mean_res_fleet << endl; report << " " << endl; report << " std deviation of fleet residuals " << endl; report << sd_res_fleet << endl;

153

report <<" Fishing mortality estimates "<< endl; report <<" "<< endl; report <<" F by age and fishery "<< endl; report << f_by_age << endl; report <<" total F by fishery "<< endl; report << tot_f << endl; report <<" total Z by age by fishery "<< endl; report << z_at_age << endl; report <<" ssb by year "<< endl; report << SSB << endl; report <<" likelihood terms "<< endl;\ report <<" "<< endl; report <<" index likelihood "<< endl; report << lik_ndx << endl; report <<" catch likelihood "<< endl; report << lik_catch << endl; report <<" ssb ratio likelihood "<< endl; report << lik_ssbc << endl; report <<" recruitment deviation likelihood "<< endl; report << lik_recc << endl; report << " " <<endl; report << " " <<endl; report <<" total likelihood "<< endl; report << f << endl; report <<" blank line "<< endl; report << "blank line " << endl; report <<" blank line "<< endl; report <<" blank line "<< endl; report << "blank line "<< endl; report << " reference points " << endl; report <<" MSY "<< endl; report << msy_out << endl; report <<" F MSY "<< endl; report << F_msy_out << endl; report <<" SSB MSY "<< endl; report << SSB_msy_out << endl; report <<" sel_msy "<< endl;

154

report << sel_msy << endl; report <<" average landings in last 3 yrs "<< endl; report << avg_land << endl; report << " " <<endl; report << "catch at age" << endl; report << c_at_age << endl; report << " " <<endl; report << "f by yr avg 1-11" << endl; report << f_by_yr << endl; /////////////////////////////////////////////////////////////////////////////////////